MIT Latest News

Subscribe to MIT Latest News feed
MIT News is dedicated to communicating to the media and the public the news and achievements of the students, faculty, staff and the greater MIT community.
Updated: 12 hours 2 min ago

AI generates high-quality images 30 times faster in a single step

Thu, 03/21/2024 - 9:30am

In our current age of artificial intelligence, computers can generate their own “art” by way of diffusion models, iteratively adding structure to a noisy initial state until a clear image or video emerges. Diffusion models have suddenly grabbed a seat at everyone’s table: Enter a few words and experience instantaneous, dopamine-spiking dreamscapes at the intersection of reality and fantasy. Behind the scenes, it involves a complex, time-intensive process requiring numerous iterations for the algorithm to perfect the image.

MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) researchers have introduced a new framework that simplifies the multi-step process of traditional diffusion models into a single step, addressing previous limitations. This is done through a type of teacher-student model: teaching a new computer model to mimic the behavior of more complicated, original models that generate images. The approach, known as distribution matching distillation (DMD), retains the quality of the generated images and allows for much faster generation. 

“Our work is a novel method that accelerates current diffusion models such as Stable Diffusion and DALLE-3 by 30 times,” says Tianwei Yin, an MIT PhD student in electrical engineering and computer science, CSAIL affiliate, and the lead researcher on the DMD framework. “This advancement not only significantly reduces computational time but also retains, if not surpasses, the quality of the generated visual content. Theoretically, the approach marries the principles of generative adversarial networks (GANs) with those of diffusion models, achieving visual content generation in a single step — a stark contrast to the hundred steps of iterative refinement required by current diffusion models. It could potentially be a new generative modeling method that excels in speed and quality.”

This single-step diffusion model could enhance design tools, enabling quicker content creation and potentially supporting advancements in drug discovery and 3D modeling, where promptness and efficacy are key.

Distribution dreams

DMD cleverly has two components. First, it uses a regression loss, which anchors the mapping to ensure a coarse organization of the space of images to make training more stable. Next, it uses a distribution matching loss, which ensures that the probability to generate a given image with the student model corresponds to its real-world occurrence frequency. To do this, it leverages two diffusion models that act as guides, helping the system understand the difference between real and generated images and making training the speedy one-step generator possible.

The system achieves faster generation by training a new network to minimize the distribution divergence between its generated images and those from the training dataset used by traditional diffusion models. “Our key insight is to approximate gradients that guide the improvement of the new model using two diffusion models,” says Yin. “In this way, we distill the knowledge of the original, more complex model into the simpler, faster one, while bypassing the notorious instability and mode collapse issues in GANs.” 

Yin and colleagues used pre-trained networks for the new student model, simplifying the process. By copying and fine-tuning parameters from the original models, the team achieved fast training convergence of the new model, which is capable of producing high-quality images with the same architectural foundation. “This enables combining with other system optimizations based on the original architecture to further accelerate the creation process,” adds Yin. 

When put to the test against the usual methods, using a wide range of benchmarks, DMD showed consistent performance. On the popular benchmark of generating images based on specific classes on ImageNet, DMD is the first one-step diffusion technique that churns out pictures pretty much on par with those from the original, more complex models, rocking a super-close Fréchet inception distance (FID) score of just 0.3, which is impressive, since FID is all about judging the quality and diversity of generated images. Furthermore, DMD excels in industrial-scale text-to-image generation and achieves state-of-the-art one-step generation performance. There's still a slight quality gap when tackling trickier text-to-image applications, suggesting there's a bit of room for improvement down the line. 

Additionally, the performance of the DMD-generated images is intrinsically linked to the capabilities of the teacher model used during the distillation process. In the current form, which uses Stable Diffusion v1.5 as the teacher model, the student inherits limitations such as rendering detailed depictions of text and small faces, suggesting that DMD-generated images could be further enhanced by more advanced teacher models. 

“Decreasing the number of iterations has been the Holy Grail in diffusion models since their inception,” says Fredo Durand, MIT professor of electrical engineering and computer science, CSAIL principal investigator, and a lead author on the paper. “We are very excited to finally enable single-step image generation, which will dramatically reduce compute costs and accelerate the process.” 

“Finally, a paper that successfully combines the versatility and high visual quality of diffusion models with the real-time performance of GANs,” says Alexei Efros, a professor of electrical engineering and computer science at the University of California at Berkeley who was not involved in this study. “I expect this work to open up fantastic possibilities for high-quality real-time visual editing.” 

Yin and Durand’s fellow authors are MIT electrical engineering and computer science professor and CSAIL principal investigator William T. Freeman, as well as Adobe research scientists Michaël Gharbi SM '15, PhD '18; Richard Zhang; Eli Shechtman; and Taesung Park. Their work was supported, in part, by U.S. National Science Foundation grants (including one for the Institute for Artificial Intelligence and Fundamental Interactions), the Singapore Defense Science and Technology Agency, and by funding from Gwangju Institute of Science and Technology and Amazon. Their work will be presented at the Conference on Computer Vision and Pattern Recognition in June.

Optimizing nuclear fuels for next-generation reactors

Wed, 03/20/2024 - 2:45pm

In 2010, when Ericmoore Jossou was attending college in northern Nigeria, the lights would flicker in and out all day, sometimes lasting only for a couple of hours at a time. The frustrating experience reaffirmed Jossou’s realization that the country’s sporadic energy supply was a problem. It was the beginning of his path toward nuclear engineering.

Because of the energy crisis, “I told myself I was going to find myself in a career that allows me to develop energy technologies that can easily be scaled to meet the energy needs of the world, including my own country,” says Jossou, an assistant professor in a shared position between the departments of Nuclear Science and Engineering (NSE), where is the John Clark Hardwick (1986) Professor, and of Electrical Engineering and Computer Science.

Today, Jossou uses computer simulations for rational materials design, AI-aided purposeful development of cladding materials and fuels for next-generation nuclear reactors. As one of the shared faculty hires between the MIT Schwarzman College of Computing and departments across MIT, his appointment recognizes his commitment to computing for climate and the environment.

A well-rounded education in Nigeria

Growing up in Lagos, Jossou knew education was about more than just bookish knowledge, so he was eager to travel and experience other cultures. He would start in his own backyard by traveling across the Niger river and enrolling in Ahmadu Bello University in northern Nigeria. Moving from the south was a cultural education with a different language and different foods. It was here that Jossou got to try and love tuwo shinkafa, a northern Nigerian rice-based specialty, for the first time.

After his undergraduate studies, armed with a bachelor’s degree in chemistry, Jossou was among a small cohort selected for a specialty master’s training program funded by the World Bank Institute and African Development Bank. The program at the African University of Science and Technology in Abuja, Nigeria, is a pan-African venture dedicated to nurturing homegrown science talent on the continent. Visiting professors from around the world taught intensive three-week courses, an experience which felt like drinking from a fire hose. The program widened Jossou’s views and he set his sights on a doctoral program with an emphasis on clean energy systems.

A pivot to nuclear science

While in Nigeria, Jossou learned of Professor Jerzy Szpunar at the University of Saskatchewan in Canada, who was looking for a student researcher to explore fuels and alloys for nuclear reactors. Before then, Jossou was lukewarm on nuclear energy, but the research sounded fascinating. The Fukushima, Japan, incident was recently in the rearview mirror and Jossou remembered his early determination to address his own country’s energy crisis. He was sold on the idea and graduated with a doctoral degree from the University of Saskatchewan on an international dean’s scholarship.

Jossou’s postdoctoral work registered a brief stint at Brookhaven National Laboratory as staff scientist. He leaped at the opportunity to join MIT NSE as a way of realizing his research interest and teaching future engineers. “I would really like to conduct cutting-edge research in nuclear materials design and to pass on my knowledge to the next generation of scientists and engineers and there’s no better place to do that than at MIT,” Jossou says.

Merging material science and computational modeling

Jossou’s doctoral work on designing nuclear fuels for next-generation reactors forms the basis of research his lab is pursuing at MIT NSE. Nuclear reactors that were built in the 1950s and ’60s are getting a makeover in terms of improved accident tolerance. Reactors are not confined to one kind, either: We have micro reactors and are now considering ones using metallic nuclear fuels, Jossou points out. The diversity of options is enough to keep researchers busy testing materials fit for cladding, the lining that prevents corrosion of the fuel and release of radioactive fission products into the surrounding reactor coolant.

The team is also investigating fuels that improve burn-up efficiencies, so they can last longer in the reactor. An intriguing approach has been to immobilize the gas bubbles that arise from the fission process, so they don’t grow and degrade the fuel.

Since joining MIT in July 2023, Jossou is setting up a lab that optimizes the composition of accident-tolerant nuclear fuels. He is leaning on his materials science background and looping computer simulations and artificial intelligence in the mix.

Computer simulations allow the researchers to narrow down the potential field of candidates, optimized for specific parameters, so they can synthesize only the most promising candidates in the lab. And AI’s predictive capabilities guide researchers on which materials composition to consider next. “We no longer depend on serendipity to choose our materials, our lab is based on rational materials design,” Jossou says, “we can rapidly design advanced nuclear fuels.”

Advancing energy causes in Africa

Now that he is at MIT, Jossou admits the view from the outside is different. He now harbors a different perspective on what Africa needs to address some of its challenges. “The starting point to solve our problems is not money; it needs to start with ideas,” he says, “we need to find highly skilled people who can actually solve problems.” That job involves adding economic value to the rich arrays of raw materials that the continent is blessed with. It frustrates Jossou that Niger, a country rich in raw material for uranium, has no nuclear reactors of its own. It ships most of its ore to France. “The path forward is to find a way to refine these materials in Africa and to be able to power the industries on that continent as well,” Jossou says.

Jossou is determined to do his part to eliminate these roadblocks.

Anchored in mentorship, Jossou’s solution aims to train talent from Africa in his own lab. He has applied for a MIT Global Experiences MISTI grant to facilitate travel and research studies for Ghanaian scientists. “The goal is to conduct research in our facility and perhaps add value to indigenous materials,” Jossou says.

Adding value has been a consistent theme of Jossou’s career. He remembers wanting to become a neurosurgeon after reading “Gifted Hands,” moved by the personal story of the author, Ben Carson. As Jossou grew older, however, he realized that becoming a doctor wasn’t necessarily what he wanted. Instead, he was looking to add value. “What I wanted was really to take on a career that allows me to solve a societal problem.” The societal problem of clean and safe energy for all is precisely what Jossou is working on today.

Study: Life’s building blocks are surprisingly stable in Venus-like conditions

Wed, 03/20/2024 - 12:00am

If there is life in the solar system beyond Earth, it might be found in the clouds of Venus. In contrast to the planet’s blisteringly inhospitable surface, Venus’ cloud layer, which extends from 30 to 40 miles above the surface, hosts milder temperatures that could support some extreme forms of life.

If it’s out there, scientists have assumed that any Venusian cloud inhabitant would look very different from life forms on Earth. That’s because the clouds themselves are made from highly toxic droplets of sulfuric acid — an intensely corrosive chemical that is known to dissolve metals and destroy most biological molecules on Earth.

But a new study by MIT researchers may challenge that assumption. Appearing today in the journal Astrobiology, the study reports that, in fact, some key building blocks of life can persist in solutions of concentrated sulfuric acid.

The study’s authors have found that 19 amino acids that are essential to life on Earth are stable for up to four weeks when placed in vials of sulfuric acid at concentrations similar to those in Venus’ clouds. In particular, they found that the molecular “backbone” of all 19 amino acids remained intact in sulfuric acid solutions ranging in concentration from 81 to 98 percent.  

“What is absolutely surprising is that concentrated sulfuric acid is not a solvent that is universally hostile to organic chemistry,” says study co-author Janusz Petkowski, a research affiliate in MIT’s Department of Earth, Atmospheric and Planetary Sciences (EAPS).

“We are finding that building blocks of life on Earth are stable in sulfuric acid, and this is very intriguing for the idea of the possibility of life on Venus,” adds study author Sara Seager, MIT’s Class of 1941 Professor of Planetary Sciences in EAPS and a professor in the departments of Physics and of Aeronautics and Astronautics. “It doesn’t mean that life there will be the same as here. In fact, we know it can’t be. But this work advances the notion that Venus’ clouds could support complex chemicals needed for life.”

The study’s co-authors include first author Maxwell Seager, an undergraduate in the Department of Chemistry at Worcester Polytechnic Institute and Seager’s son, and William Bains, a research affiliate at MIT and a scientist at Cardiff University.

Building blocks in acid

The search for life in Venus’ clouds has gained momentum in recent years, spurred in part by a controversial detection of phosphine — a molecule that is considered to be one signature of life — in the planet’s atmosphere. While that detection remains under debate, the news has reinvigorated an old question: Could Earth’s sister planet actually host life?

In search of an answer, scientists are planning several missions to Venus, including the first largely privately funded mission to the planet, backed by California-based launch company Rocket Lab. That mission, on which Seager is the science principal investigator, aims to send a spacecraft through the planet’s clouds to analyze their chemistry for signs of organic molecules.

Ahead of the mission’s January 2025 launch, Seager and her colleagues have been testing various molecules in concentrated sulfuric acid to see what fragments of life on Earth might also be stable in Venus’ clouds, which are estimated to be orders of magnitude more acidic than the most acidic places on Earth.

“People have this perception that concentrated sulfuric acid is an extremely aggressive solvent that will chop everything to pieces,” Petkowski says. “But we are finding this is not necessarily true.”

In fact, the team has previously shown that complex organic molecules such as some fatty acids and nucleic acids remain surprisingly stable in sulfuric acid. The scientists are careful to emphasize, as they do in their current paper, that “complex organic chemistry is of course not life, but there is no life without it.”

In other words, if certain molecules can persist in sulfuric acid, then perhaps the highly acidic clouds of Venus are habitable, if not necessarily inhabited.

In their new study, the team turned their focus on amino acids — molecules that combine to make essential proteins, each with their own specific function. Every living thing on Earth requires amino acids to make proteins that in turn carry out life-sustaining functions, from breaking down food to generating energy, building muscle, and repairing tissue.

“If you consider the four major building blocks of life as nucleic acid bases, amino acids, fatty acids, and carbohydrates, we have demonstrated that some fatty acids can form micelles and vesicles in sulfuric acid, and the nucleic acid bases are stable in sulfuric acid. Carbohydrates have been shown to be highly reactive in sulfuric acid,” Maxwell
Seager explains. “That only left us with amino acids as the last major building block to
study.”

A stable backbone

The scientists began their studies of sulfuric acid during the pandemic, carrying out their experiments in a home laboratory. Since that time, Seager and her son continued work on chemistry in concentrated sulfuric acid. In early 2023, they ordered powder samples of 20 “biogenic” amino acids — those amino acids that are essential to all life on Earth. They dissolved each type of amino acid in vials of sulfuric acid mixed with water, at concentrations of 81 and 98 percent, which represent the range that exists in Venus’ clouds.

The team then let the vials incubate for a day before transporting them to MIT’s Department of Chemistry Instrumentation Facility (DCIF), a shared, 24/7 laboratory that offers a number of automated and manual instruments for MIT scientists to use. For their part, Seager and her team used the lab’s nuclear magnetic resonance (NMR) spectrometer to analyze the structure of amino acids in sulfuric acid.

After analyzing each vial several times over four weeks, the scientists found, to their surprise, that the basic molecular structure, or “backbone” in 19 of the 20 amino acids remained stable and unchanged, even in highly acidic conditions.

“Just showing that this backbone is stable in sulfuric acid doesn’t mean there is life on Venus,” notes Maxwell Seager. “But if we had shown that this backbone was compromised, then there would be no chance of life as we know it.”

“Now, with the discovery that many amino acids and nucleic acids are stable in 98 percent sulfuric acid, the possibility of life surviving in sulfuric acid may not be so far-fetched or fantastic,” says Sanjay Limaye, a planetary scientist at the University of Wisconsin who has studied Venus for over 45 years, and who was not involved with this study. “Of course, many obstacles lie ahead, but life that evolved in water and adapted to sulfuric acid may not be easily dismissed.”

The team acknowledges that Venus’ cloud chemistry is likely messier than the study’s “test tube” conditions. For instance, scientists have measured various trace gases, in addition to sulfuric acid, in the planet’s clouds. As such, the team plans to incorporate certain trace gases in future experiments.

“There are only a few groups in the world now that are working on chemistry in sulfuric acid, and they will all agree that no one has intuition,” adds Sara Seager. “I think we are just more happy than anything that this latest result adds one more ‘yes’ for the possibility of life on Venus.”

Creative collisions: Crossing the art-science divide

Tue, 03/19/2024 - 3:40pm

MIT has a rich history of productive collaboration between the arts and the sciences, anchored by the conviction that these two conventionally opposed ways of thinking can form a deeply generative symbiosis that serves to advance and humanize new technologies. 

This ethos was made tangible when the Bauhaus artist and educator György Kepes established the MIT Center for Advanced Visual Studies (CAVS) within the Department of Architecture in 1967. CAVS has since evolved into the Art, Culture, and Technology (ACT) program, which fosters close links to multiple other programs, centers, and labs at MIT. Class 4.373/4.374 (Creating Art, Thinking Science), open to undergraduates and master’s students of all disciplines as well as certain students from the Harvard Graduate School of Design (GSD), is one of the program’s most innovative offerings, proposing a model for how the relationship between art and science might play out at a time of exponential technological growth. 

Now in its third year, the class is supported by an Interdisciplinary Class Development Grant from the MIT Center for Art, Science and Technology (CAST) and draws upon the unparalleled resources of MIT.nano; an artist’s high-tech toolbox for investigating the hidden structures and beauty of our material universe.

High ambitions and critical thinking

The class was initiated by Tobias Putrih, lecturer in ACT, and is taught with the assistance of Ardalan SadeghiKivi MArch ’23, and Aubrie James SM ’24. Central to the success of the class has been the collaboration with co-instructor Vladimir Bulović, the founding director of MIT.nano and Fariborz Maseeh Chair in Emerging Technology, who has positioned the facility as an open-access resource for the campus at large — including MIT’s community of artists. “Creating Art, Thinking Science” unfolds the 100,000 square feet of cleanroom and lab space within the Lisa T. Su Building, inviting participating students to take advantage of cutting-edge equipment for nanoscale visualization and fabrication; in the hands of artists, devices for discovering nanostructures and manipulating atoms become tools for rendering the invisible visible and deconstructing the dynamics of perception itself. 

The expansive goals of the class are tempered by an in-built criticality. “ACT has a unique position as an art program nested within a huge scientific institute — and the challenges of that partnership should not be underestimated,” reflects Putrih. “Science and art are wholly different knowledge systems with distinct historical perspectives. So, how do we communicate? How do we locate that middle ground, that third space?”

An evolving answer, tested and developed throughout the partnership between ACT and MIT.nano, involves a combination of attentive mentorship and sharing of artistic ideas, combined with access to advanced technological resources and hands-on practical training. 

“MIT.nano currently accommodates more than 1,200 individuals to do their work, across 250 different research groups,” says Bulović. “The fact that we count artists among those is a matter of pride for us. We’ve found that the work of our scientists and technologists is enhanced by having access to the language of art as a form of expression — equally, the way that artists express themselves can be stretched beyond what could previously be imagined, simply by having access to the tools and instruments at MIT.nano.”

A playground for experimentation

True to the spirit of the scientific method and artistic iteration, the class is envisioned as a work in progress — a series of propositions and prototypes for how dialogue between scientists and artists might work in practice. The outcomes of those experiments can now be seen installed in the first and second floor galleries at MIT.nano. As part of the facility’s five-year anniversary celebration, the class premiered an exhibition showcasing works created during previous years of “Creating Art, Thinking Science.” 

Visitors to the exhibition, “zero.zerozerozerozerozerozerozerozeroone” (named for the numerical notation for one nanometer), will encounter artworks ranging from a minimalist silicon wafer produced with two-photon polymerization (2PP) technology (“Obscured Invisibility,” 2021, Hyun Woo Park), to traces of an attempt to make vegetable soup in the cleanroom using equipment such as a cryostat, a fluorescing microscope, and a Micro-CT scanner (“May I Please Make You Some Soup?,” 2022, Simone Lasser). 

These works set a precedent for the artworks produced during the fall 2023 iteration of the class. For Ryan Yang, in his senior year studying electrical engineering and computer science at MIT, the chance to engage in open discussion and experimental making has been a rare opportunity to “try something that might not work.” His project explores the possibilities of translating traditional block printing techniques to micron-scale 3D-printing in the MIT.nano labs.

Yang has taken advantage of the arts curriculum at MIT at an early stage in his academic career as an engineer; meanwhile, Ameen Kaleem started out as a filmmaker in New Delhi and is now pursuing a master’s degree in design engineering at Harvard GSD, cross-registered at MIT. 

Kaleem’s project models the process of abiogenesis (the evolution of living organisms from inorganic or inanimate substances) by bringing living moss into the MIT.nano cleanroom facilities to be examined at an atomic scale. “I was interested in the idea that, as a human being in the cleanroom, you are both the most sanitized version of yourself and the dirtiest thing in that space,” she reflects. “Drawing attention to the presence of organic life in the cleanroom is comparable to bringing art into spaces where it might not otherwise exist — a way of humanizing scientific and technological endeavors.”

Consciousness, immersion, and innovation

The students draw upon the legacies of landmark art-science initiatives — including international exhibitions such as “Cybernetic Serendipity” (London ICA, 1968), the “New Tendencies” series (Zagreb, 1961-73), and “Laboratorium” (Antwerp, 1999) — and take inspiration from the instructors’ own creative investigations of the inner workings of different knowledge systems. “In contemporary life, and at MIT in particular, we’re immersed in technology,” says Putrih. “It’s the nature of art to reveal that to us, so that we might see the implications of what we are producing and its potential impact.”

By fostering a mindset of imagination and criticality, combined with building the technical skills to address practical problems, “Creating Art, Thinking Science” seeks to create the conditions for a more expansive version of technological optimism; a culture of innovation in which social and environmental responsibility are seen as productive parameters for enriched creativity. The ripple effects of the class might be years in the making, but as Bulović observes while navigating the exhibition at MIT.nano, “The joy of the collaboration can be felt in the artworks.”

Visiting scholars from Ukraine kick off Global MIT At-Risk Fellows Program

Tue, 03/19/2024 - 3:15pm

Even before Russia’s invasion of Ukraine two years ago, members of the MIT faculty knew that violence and political pressures in the region endangered the work and well-being of Ukrainian scholars and contemplated how MIT could assist. The start of the full-scale invasion in February 2022 was the decisive catalyst — triggering the launch of the MIT-Ukraine Program later that year and eventually spurring creation of the new Global MIT At-Risk Fellows (GMAF) program with an initial focus on Ukraine.

Designed to provide sanctuary to scholars around the globe whose lives and academic freedom have been upended by war and tragedy in their countries, GMAF aspires to bring up to five international scholars annually to the MIT campus for semester-long study and research that will ultimately benefit their countries and simultaneously enrich the MIT community. Welcoming the program’s first three visiting scholars from Ukraine, GMAF officially kicked off on Feb. 29 at a reception hosted by the Office of the Vice Provost for International Activities and the MIT Center for International Studies.

The reception showcased the varied struggles of displaced individuals with the photographic exhibition, “Standing for freedom, portraits of scientists in exile,” comprising portraits of refugee scholars from countries torn by war and political upheaval. This inaugural U.S. installation will be on public display at MIT’s Koch Institute Public Galleries (Building 76) from April 3 through April 30. It then travels to the French Embassy in Washington. It is the work of PAUSE, a French organization that has enabled scientists in exile to continue their work in France since 2017.

“It’s the first time the exhibit has been in the United States, and we are very proud and honored that it is here,” says PAUSE Executive Director Laura Loheac, who participated in the Feb. 29 event along with PAUSE co-founder Professor Pascale Laborier, photographer Pierre-Jérôme Adjedj, members of the local Ukrainian community, and MIT faculty, students, and senior staff.

Ford International Professor of History Elizabeth Wood said Russia’s full-scale invasion of Ukraine “is not only tragic in its own right,” but “has also created a host of dire scientific and technological problems that we think MIT faculty, staff, and students are well positioned to help solve in collaboration with Ukrainians themselves.”

“Our focus in the MIT-Ukraine Program — itself launched just 16 months ago — has been to serve as a Ukraine hub at MIT,” said Wood, faculty chair for both GMAF and MIT-Ukraine. “The core idea of the GMAF Program in its current incarnation is to bring Ukrainian scholars to MIT for a semester so they can have a bit of a refuge from the war — though I know it is never far from their minds, and so they can soak up some of MIT’s famous culture of 'mens et manus' — mind, hands, and heart.”

GMAF scholars Liudmyla Huliaieva and Kateryna Lopatiuk have been at MIT for about a month, while the cohort’s third member, Dmytro Chumachenko, arrived one day before the reception due to visa processing delays. Huliaieva is an economist focused on the economic adaptation and survival of Ukrainian displaced women, while Lopatiuk is an architect and urban planner involved in rebuilding cities and towns across Ukraine, and Chumachenko is a multidisciplinary scientist working at the intersection of artificial intelligence and public health. All met rigorous criteria considered by faculty committee members who evaluated 80 applications for GMAF’s first group of scholars.

“We wanted individuals who were deeply committed to helping Ukraine, who could benefit from a place at MIT, who were providing absolutely top-notch scholarship, who could actually leave the country — since many men and some women cannot do that because of circumstances of the war — and who had projects they were ready and eager to pursue while here,” Wood says.

Huliaieva, Lopatiuk, and Chumachenko are the first of what will likely be 10 Ukrainian researchers and faculty spending a semester at MIT during the two-year GMAF pilot program. With additional funding, the program is envisioned to eventually expand to help scholars in other countries where their work is jeopardized by war or displacement. Provost Cynthia Barnhart says the three Ukrainian scholars now on campus “represent just the start.”

Event speakers noted GMAF’s collaborative nature. Among those recognized for conceiving and organizing it were MIT Vice Provost for International Activities Richard Lester, Senior Director Beth Dupuy, and Institute Professor Suzanne Berger event emcee and founding director of the MIT International Science and Technology Initiatives (MISTI). Credited for implementing the new program was Svitlana Krasynska, program director for both MIT-Ukraine and GMAF.

Lester said about the program, “The threats to science and scholarship from war and political repression are profound and, unfortunately, they are growing around the world. Even though the GMAF program is small relative to the vast need, it is a practical way for MIT to contribute and also to demonstrate our solidarity with vulnerable members of the global academic community of which we are part.”

Krasynska said in an interview that, although the exact number is currently unknown, it is estimated that over 60,000 Ukrainian scholars and support staff have been displaced and many universities destroyed or badly damaged in the past two years.

“Lives have been severely disrupted,” said Krasynska, who was born and raised in Ukraine and has lived in the United States since 1997. “We really need to support Ukrainian scientists and support Ukrainian science because it is in dire straits right now.”

Chumachenko said his home campus, the National Aerospace University Kharkiv Aviation Institute, has suffered 160 Russian bombs, “but we are still working and teaching.”

“Besides what we bring back to Ukraine, I believe the three of us can bring something here,” he said. “People know about the Russian war in Ukraine through TV, but it’s not always the full picture.”

Lopatiuk echoed those sentiments. Noting that when she applied to the GMAF program she had several research goals in mind, but realized after spending the past month at MIT that “my main purpose is also to get students to get to know what Ukraine is as a country beyond the consequences of war” — including the nation’s history, culture and ideas.

Noting that her first impression of MIT “is that it’s a very big, friendly family,” Huliaieva plans to present a virtual seminar at Harvard University on March 18 designed to broaden awareness and understanding of the challenges faced by Ukrainians — both those still there and people forced to leave. Titled “Dreaming of home: Displaced Ukrainian women between transience and permanency,” it reflects her research into helping Ukrainian women gain financial independence and freedom.

Barnhart welcomed Huliaieva, Lopatiuk, and Chumachenko to MIT “not only as our very first cohort of scholars, but also as colleagues and collaborators.”

“I hope you’ll find our entire campus is a thriving ecosystem of ideas and innovation,” she said. “I hope you will learn that we are deeply committed to protecting education and scholarship whenever they come under threat.”

Visting scholars from Ukraine kick off Global MIT At-Risk Fellows Program

Tue, 03/19/2024 - 3:15pm

Even before Russia’s invasion of Ukraine two years ago, members of the MIT faculty knew that violence and political pressures in the region endangered the work and well-being of Ukrainian scholars and contemplated how MIT could assist. The start of the full-scale invasion in February 2022 was the decisive catalyst — triggering the launch of the MIT-Ukraine Program later that year and eventually spurring creation of the new Global MIT At-Risk Fellows (GMAF) program with an initial focus on Ukraine.

Designed to provide sanctuary to scholars around the globe whose lives and academic freedom have been upended by war and tragedy in their countries, GMAF aspires to bring up to five international scholars annually to the MIT campus for semester-long study and research that will ultimately benefit their countries and simultaneously enrich the MIT community. Welcoming the program’s first three visiting scholars from Ukraine, GMAF officially kicked off on Feb. 29 at a reception hosted by the Office of the Vice Provost for International Activities and the MIT Center for International Studies.

The reception showcased the varied struggles of displaced individuals with the photographic exhibition, “Standing for freedom, portraits of scientists in exile,” comprising portraits of refugee scholars from countries torn by war and political upheaval. This inaugural U.S. installation will be on public display at MIT’s Koch Institute Public Galleries (Building 76) from April 3 through April 30. It then travels to the French Embassy in Washington. It is the work of PAUSE, a French organization that has enabled scientists in exile to continue their work in France since 2017.

“It’s the first time the exhibit has been in the United States, and we are very proud and honored that it is here,” says PAUSE Executive Director Laura Loheac, who participated in the Feb. 29 event along with PAUSE co-founder Professor Pascale Laborier, photographer Pierre-Jérôme Adjedj, members of the local Ukrainian community, and MIT faculty, students, and senior staff.

Ford International Professor of History Elizabeth Wood said Russia’s full-scale invasion of Ukraine “is not only tragic in its own right,” but “has also created a host of dire scientific and technological problems that we think MIT faculty, staff, and students are well positioned to help solve in collaboration with Ukrainians themselves.”

“Our focus in the MIT-Ukraine Program — itself launched just 16 months ago — has been to serve as a Ukraine hub at MIT,” said Wood, faculty chair for both GMAF and MIT-Ukraine. “The core idea of the GMAF Program in its current incarnation is to bring Ukrainian scholars to MIT for a semester so they can have a bit of a refuge from the war — though I know it is never far from their minds, and so they can soak up some of MIT’s famous culture of 'mens et manus' — mind, hands, and heart.”

GMAF scholars Liudmyla Huliaieva and Kateryna Lopatiuk have been at MIT for about a month, while the cohort’s third member, Dmytro Chumachenko, arrived one day before the reception due to visa processing delays. Huliaieva is an economist focused on the economic adaptation and survival of Ukrainian displaced women, while Lopatiuk is an architect and urban planner involved in rebuilding cities and towns across Ukraine, and Chumachenko is a multidisciplinary scientist working at the intersection of artificial intelligence and public health. All met rigorous criteria considered by faculty committee members who evaluated 80 applications for GMAF’s first group of scholars.

“We wanted individuals who were deeply committed to helping Ukraine, who could benefit from a place at MIT, who were providing absolutely top-notch scholarship, who could actually leave the country — since many men and some women cannot do that because of circumstances of the war — and who had projects they were ready and eager to pursue while here,” Wood says.

Huliaieva, Lopatiuk, and Chumachenko are the first of what will likely be 10 Ukrainian researchers and faculty spending a semester at MIT during the two-year GMAF pilot program. With additional funding, the program is envisioned to eventually expand to help scholars in other countries where their work is jeopardized by war or displacement. Provost Cynthia Barnhart says the three Ukrainian scholars now on campus “represent just the start.”

Event speakers noted GMAF’s collaborative nature. Among those recognized for conceiving and organizing it were MIT Vice Provost for International Activities Richard Lester, Senior Director Beth Dupuy, and Institute Professor Suzanne Berger event emcee and founding director of the MIT International Science and Technology Initiatives (MISTI). Credited for implementing the new program was Svitlana Krasynska, program director for both MIT-Ukraine and GMAF.

Lester said about the program, “The threats to science and scholarship from war and political repression are profound and, unfortunately, they are growing around the world. Even though the GMAF program is small relative to the vast need, it is a practical way for MIT to contribute and also to demonstrate our solidarity with vulnerable members of the global academic community of which we are part.”

Krasynska said in an interview that, although the exact number is currently unknown, it is estimated that over 60,000 Ukrainian scholars and support staff have been displaced and many universities destroyed or badly damaged in the past two years.

“Lives have been severely disrupted,” said Krasynska, who was born and raised in Ukraine and has lived in the United States since 1997. “We really need to support Ukrainian scientists and support Ukrainian science because it is in dire straits right now.”

Chumachenko said his home campus, the National Aerospace University Kharkiv Aviation Institute, has suffered 160 Russian bombs, “but we are still working and teaching.”

“Besides what we bring back to Ukraine, I believe the three of us can bring something here,” he said. “People know about the Russian war in Ukraine through TV, but it’s not always the full picture.”

Lopatiuk echoed those sentiments. Noting that when she applied to the GMAF program she had several research goals in mind, but realized after spending the past month at MIT that “my main purpose is also to get students to get to know what Ukraine is as a country beyond the consequences of war” — including the nation’s history, culture and ideas.

Noting that her first impression of MIT “is that it’s a very big, friendly family,” Huliaieva plans to present a virtual seminar at Harvard University on March 18 designed to broaden awareness and understanding of the challenges faced by Ukrainians — both those still there and people forced to leave. Titled “Dreaming of home: Displaced Ukrainian women between transience and permanency,” it reflects her research into helping Ukrainian women gain financial independence and freedom.

Barnhart welcomed Huliaieva, Lopatiuk, and Chumachenko to MIT “not only as our very first cohort of scholars, but also as colleagues and collaborators.”

“I hope you’ll find our entire campus is a thriving ecosystem of ideas and innovation,” she said. “I hope you will learn that we are deeply committed to protecting education and scholarship whenever they come under threat.”

3 Questions: Progress on updating MIT’s undergraduate curriculum

Tue, 03/19/2024 - 2:20pm

In late February, Vice Chancellor for Undergraduate and Graduate Education Ian A. Waitz and Faculty Chair Mary Fuller announced the formation and launch of the Task Force on the MIT Undergraduate Academic Program (TFUAP). The effort fulfills a critical recommendation of the Task Force 2021 and Beyond RIC1 (Undergraduate Program) and draws upon several, prior foundational working groups some focused on the current General Institute Requirements (GIRs) and others on updating recent studies for the purposes of this review.

In this interview, task force co-chairs Adam Martin, professor of biology, and Joel Voldman, the William R. Brody Professor of Electrical Engineering and Computer Science describe the TFUAP’s goals, approach, and next steps.

Q: The charge of the task force is quite ambitious, including “reviewing the current undergraduate academic program and considering improvements with a focus on both the curriculum and pedagogy.” Can you explain your approach?

Martin: For context, it’s important to know that the undergraduate program is multifaceted and consists of many components, including majors, electives, experiential learning, and of course the GIRs arguably one of the best-known acronyms at MIT! Moreover, the GIRs include science core classes; humanities, arts, and social sciences classes; certain electives in science and engineering; and a lab requirement, each of which serves a slightly different purpose and dovetails with majors and minors in unique ways.

Some aspects of the academic program are determined by the faculty, either MIT-wide or within a particular department. Others can be customized by students, in consultation with faculty and staff advisors, from the broad array of curricular and co-curricular offerings at MIT. The task force will look holistically at all of these aspects, considering both what MIT requires of all students, and the options we make available as students chart their own paths.

As part of this holistic approach, the TFUAP will zero in on both content and pedagogy. Obviously, the content we cover is important; our goal must remain to provide undergraduates with the world-class education they expect. But how we teach is of fundamental importance, as well. The pedagogy we adopt should be inclusive, supported by research, and designed to help students not only understand what they are learning, but why they are learning it how it relates to their majors, potential careers, and their lives.

Voldman: I think your question’s description of our charge as “ambitious” is noteworthy. We feel that the task force is ambitious, too, but perhaps in a different sense from the question. That is, we believe our job is to not only think about nuts-and-bolts issues of the academic program requirements but also to consider the big picture. What are the most expansive possibilities? How can we push the envelope? That’s the MIT way, after all.

Q: The task force is building upon quite a bit of past work and benefits from some major accomplishments recommended by Task Force 2021 (TF2021). For example, how does the creation of the Undergraduate Advising Center, and in general, the desire to provide more personal and professional support to all students, fit in with the potential updates to the undergraduate curriculum?

Martin: You’re absolutely right our work benefits greatly from years of conversations focused on the undergraduate academic program, particularly in the last decade or so. These include the 2014 Task Force on the Future of Education; the 2018 Designing the First-Year Experience Class; Task Force 2021 and Beyond (TF2021); the Foundational Working Groups (part of the RIC 1 implementation) that have studied the existing MIT undergraduate program; and the Committee on the Undergraduate Program. The valuable work of these past committees and their findings will certainly inform our thought process.

In the past, groups that evaluated the undergraduate curriculum were also charged with tackling related topics, such as undergraduate advising or revamping classrooms. Taking on any one of these three issues is ambitious by any measure! What’s changed in the past decade is that advances have been made in these other critical areas, so the TFUAP can focus solely on curriculum and pedagogy. For example, thanks to recent accomplishments by TF2021 and others, we have implemented a new advising system for all undergraduates in the form of the Undergraduate Advising Center.

We envision the TFUAP being a highly collaborative process, bringing in voices across the entire Institute and beyond. We welcome input from members of the community via email at tfuap@mit.edu. We will also be reaching out to student groups, alumni, individual faculty, faculty groups, and administrative staff across the Institute to hear their perspectives.

Q: Part of what TFUAP will have to confront, no doubt, are some of the most pressing issues of our time, like the rise of computing and AI, climate change (what President Kornbluth calls an existential threat to our way of life), and the changing nature of learning (online, hybrid, etc.). How are you thinking about all of these factors?

Voldman: That is a good question! It’s early days, and our work is just beginning, but we know that these and other issues loom over all of us. For example, we are keenly aware of the influx of students into computing-related majors and classes, and we need to think deeply about the implications. Furthermore, we want a curriculum that prepares students for current and upcoming global challenges as well as changes in the technology and tools available to address those challenges. However, we can expect that our students will need to be agile and curious, lifelong learners, collaborative and compassionate teammates, and creative and thoughtful problem-solvers.

As we work with the community to design the next version of an MIT undergraduate education, it will be important to build a structure that can incorporate the biggest challenges and opportunities of the day, while staying flexible and responsive to an ever-evolving world.

A protein found in human sweat may protect against Lyme disease

Tue, 03/19/2024 - 6:00am

Lyme disease, a bacterial infection transmitted by ticks, affects nearly half a million people in the United States every year. In most cases, antibiotics effectively clear the infection, but for some patients, symptoms linger for months or years.

Researchers at MIT and the University of Helsinki have now discovered that human sweat contains a protein that can protect against Lyme disease. They also found that about one-third of the population carries a genetic variant of this protein that is associated with Lyme disease in genome-wide association studies.

It’s unknown exactly how the protein inhibits the growth of the bacteria that cause Lyme disease, but the researchers hope to harness the protein’s protective abilities to create skin creams that could help prevent the disease, or to treat infections that don’t respond to antibiotics.

“This protein may provide some protection from Lyme disease, and we think there are real implications here for a preventative and possibly a therapeutic based on this protein,” says Michal Caspi Tal, a principal research scientist in MIT’s Department of Biological Engineering and one of the senior authors of the new study.

Hanna Ollila, a senior researcher at the Institute for Molecular Medicine at the University of Helsinki and a researcher at the Broad Institute of MIT and Harvard, is also a senior author of the paper, which appears today in Nature Communications. The paper’s lead author is Satu Strausz, a postdoc at the Institute for Molecular Medicine at the University of Helsinki.

A surprising link

Lyme disease is most often caused by a bacterium called Borrelia burgdorferi. In the United States, this bacterium is spread by ticks that are carried by mice, deer, and other animals. Symptoms include fever, headache, fatigue, and a distinctive bulls-eye rash.

Most patients receive doxycycline, an antibiotic that usually clears up the infection. In some patients, however, symptoms such as fatigue, memory problems, sleep disruption, and body aches can persist for months or years.

Tal and Ollila, who were postdocs together at Stanford University, began this study a few years ago in hopes of finding genetic markers of susceptibility to Lyme disease. To that end, they decided to run a genome-wide association study (GWAS) on a Finnish dataset that contains genome sequences for 410,000 people, along with detailed information on their medical histories.

This dataset includes about 7,000 people who had been diagnosed with Lyme disease, allowing the researchers to look for genetic variants that were more frequently found in people who had had Lyme disease, compared with those who hadn’t.

This analysis revealed three hits, including two found in immune molecules that had been previously linked with Lyme disease. However, their third hit was a complete surprise — a secretoglobin called SCGB1D2.

Secretoglobins are a family of proteins found in tissues that line the lungs and other organs, where they play a role in immune responses to infection. The researchers discovered that this particular secretoglobin is produced primarily by cells in the sweat glands.

To find out how this protein might influence Lyme disease, the researchers created normal and mutated versions of SCGB1D2 and exposed them to Borrelia burgdorferi grown in the lab. They found that the normal version of the protein significantly inhibited the growth of Borrelia burgdorferi. However, when they exposed bacteria to the mutated version, twice as much protein was required to suppress bacterial growth.

The researchers then exposed bacteria to either the normal or mutated variant of SCGB1D2 and injected them into mice. Mice injected with the bacteria exposed to the mutant protein became infected with Lyme disease, but mice injected with bacteria exposed to the normal version of SCGB1D2 did not.

“In the paper we show they stayed healthy until day 10, but we followed the mice for over a month, and they never got infected. This wasn’t a delay, this was a full stop. That was really exciting,” Tal says.

Preventing infection

After the MIT and University of Helsinki researchers posted their initial findings on a preprint server, researchers in Estonia replicated the results of the genome-wide association study, using data from the Estonian Biobank. These data, from about 210,000 people, including 18,000 with Lyme disease, were later added to the final Nature Communications study.

The researchers aren’t sure yet how SCGB1D2 inhibits bacterial growth, or why the variant is less effective. However, they did find that the variant causes a shift from the amino acid proline to leucine, which may interfere with the formation of a helix found in the normal version.

They now plan to investigate whether applying the protein to the skin of mice, which do not naturally produce SCGB1D2, could prevent them from being infected by Borrelia burgdorferi. They also plan to explore the protein’s potential as a treatment for infections that don’t respond to antibiotics.

“We have fantastic antibiotics that work for 90 percent of people, but in the 40 years we’ve known about Lyme disease, we have not budged that,” Tal says. “Ten percent of people don’t recover after having antibiotics, and there’s no treatment for them.”

“This finding opens the door to a completely new approach to preventing Lyme disease in the first place, and it will be interesting to see if it could be useful for preventing other types of skin infections too,” says Kara Spiller, a professor of biomedical innovation in the School of Biomedical Engineering at Drexel University, who was not involved in the study.

The researchers note that people who have the protective version of SCGB1D2 can still develop Lyme disease, and they should not assume that they won’t. One factor that may play a role is whether the person happens to be sweating when they’re bitten by a tick carrying Borrelia burgdorferi.

SCGB1D2 is just one of 11 secretoglobin proteins produced by the human body, and Tal also plans to study what some of those other secretoglobins may be doing in the body, especially in the lungs, where many of them are found.

“The thing I’m most excited about is this idea that secretoglobins might be a class of antimicrobial proteins that we haven’t thought about. As immunologists, we talk nonstop about immunoglobulins, but I had never heard of a secretoglobin before this popped up in our GWAS study. This is why it’s so fun for me now. I want to know what they all do,” she says.

The research was funded, in part, by Emily and Malcolm Fairbairn, the Instrumentarium Science Foundation, the Academy of Finland, the Finnish Medical Foundation, the Younger Family, and the Bay Area Lyme Foundation.

Pushing material boundaries for better electronics

Tue, 03/19/2024 - 12:00am

Undergrads, take note: The lessons you learn in those intro classes could be the key to making your next big discovery. At least, that’s been the case for MIT’s Jeehwan Kim.

A recently tenured faculty member in MIT’s departments of Mechanical Engineering and Materials Science and Engineering, Kim has made numerous discoveries about the nanostructure of materials and is funneling them directly into the advancement of next-generation electronics.

His research aims to push electronics past the inherent limits of silicon — a material that has reliably powered transistors and most other electronic elements but is reaching a performance limit as more computing power is packed into ever smaller devices.

Today, Kim and his students at MIT are exploring materials, devices, and systems that could take over where silicon leaves off. Kim is applying his insights to design next-generation devices, including low-power, high-performance transistors and memory devices, artificial intelligence chips, ultra-high-definition micro-LED displays, and flexible electronic “skin.” Ultimately, he envisions such beyond-silicon devices could be built into supercomputers small enough to fit in your pocket.

The innovations that have come out of his research are recorded in more than 200 issued U.S. patents and 70 research papers — an extensive list that he and his students continue to grow.

Kim credits many of his breakthroughs to the fundamentals he learned in his university days. In fact, he has carried his college textbooks and notes with him with every move. Today, he keeps the undergraduate notes — written in a light and meticulous graphite and ink — on a shelf nearest to his MIT desk, close at hand. He references them in his own class lectures and presentations, and when brainstorming research solutions.

“These textbooks are all in my brain now,” Kim says. “I’ve learned that if you completely understand the fundamentals, you can solve any problem.”

Fundamental shift

Kim wasn’t always a model student. Growing up in Seoul, South Korea, he was fixed on a musical career. He had a passion for singing and was bored by most other high school subjects.

“It was very monotonic,” Kim recalls. “My motivation for high school subjects was very low.”

After graduating high school, he enrolled in a materials science program at Hongik University, where he was lucky to met professors who had graduated from MIT and who later motivated him to study in the United States. But, Kim spent his first year there trying to make it as a musician. He wrote and sang songs that he recorded and sent to promoters, and went to multiple auditions. But after a year, he was faced with no call-backs, and a hard question.

“What should I do? It was a crisis to me,” Kim says.

In his second year, he decided to give materials science a go. When he sat in on his first class, he was surprised to find that the subject — the structure and behavior of materials at the atomic scale — made him want to learn more.

“My first year, my GPA was almost zero because I didn’t attend class, and was going to be kicked out,” Kim says. “Then from my second year on, I really loved every single subject in materials science. People who saw me in the library were surprised: ‘What are you doing here, without a guitar?’ I must have read these textbooks more than 10 times, and felt I really understood everything fundamental.”

Back to basics

He took this newfound passion to Seoul National University, where he enrolled in the materials science master’s program and learned to apply the ideas he absorbed to hands-on research problems. Metallurgy was a dominant field at the time, and Kim was assigned to experiment with high-temperature alloys — mixing and melting metallic powders to create materials that could be used in high-performance engines.

After completing his master’s, Kim wanted to continue with a PhD, overseas. But to do so, he first had to serve in the military. He spent the next two and a half years in the Korean air force, helping to maintain and refuel aircraft, and inventory their parts. All the while, he prepared applications to graduate schools abroad.

In 2003, after completing his service, he headed overseas, where he was accepted to the materials science graduate program at the University of California at Los Angeles with a fellowship.

“When I came out of the airplane and went to the dorm for the first day, people were drinking Corona on the balcony, playing music, and there was beautiful weather, and I thought, this is where I’m supposed to be!” Kim recalls.

For his PhD, he began to dive into the microscopic world of electronic materials, seeking ways to manipulate them to make faster electronics. The subject was a focus for his advisor, who previously worked at Bell Labs, where many computing innovations originated at the time.

“A lot of the papers I was reading were from Bell Labs, and IBM T.J. Watson, and I was so impressed, and thought: I really want to be a scientist there. That was my dream,” Kim says.

During his PhD program, he reached out to a scientist at IBM whose name kept coming up in the papers Kim was reading. In his initial letter, Kim wrote with a question about his own PhD work, which tackled a hard industry problem: how to stretch, or “strain,” silicon to minimize defects that would occur as more transistors are packed on a chip. 

The query opened a dialogue, and Kim eventually inquired and was accepted to an internship at the IBM T.J. Watson Research Center, just outside New York City. Soon after he arrived, his manager pitched him a challenge: He might be hired full-time if he could solve a new, harder problem, having to do with replacing silicon.

At the time, the electronics industry was looking to germanium as a possible successor to silicon. The material can conduct electrons at even smaller scales, which would enable germanium to be made into even tinier transistors, for faster, smaller, and more powerful devices. But there was no reliable way for germanium to be “doped” — an essential process that replaces some of a material’s atoms with another type of atom in a way that controls how electrons flow through the material.

“My manager told me he didn’t expect me to solve this. But I really wanted the job,” Kim says. “So day and night, I thought, how to solve this? And I always went back to the textbooks.”

Those textbooks reminded him of a fundamental rule: Replacing one atom with another would work well if both atoms were of similar size. This revelation triggered an idea. Perhaps germanium could be doped with a combination of two different atoms with an average atomic size that is similar to germanium’s.

“I came up with this idea, and right after, IBM showed that it worked. I was so amazed,” Kim says. “From that point, research became my passion. I did it because it was just so fun. Singing is not so different from performing research.”

As promised, he was hired as a postdoc and soon after, promoted to research staff member — a title he carried, literally, with pride.

“I was feeling so happy to be there,” Kim says. “I even wore my IBM badge to restaurants, and everywhere I went.”

Throughout his time at IBM, he learned to focus on research that directly impacts everyday human life, and how to apply the fundamentals to develop next-generation products.

“IBM really raised me up as an engineer who can identify the problems in an industry and find creative solutions to tackle the challenges,” he says.

Cycle of life

And yet, Kim felt he could do more. He was working on boundary-pushing research at one of the leading innovation hubs in the country, where “out-of-the-box” thinking was encouraged, and experimentally tested. But he wanted to explore beyond the company’s research portfolio, and also, find a way to pursue research not just as a profession but as a passion.

“My experience taught me that you can lead a very happy life as an engineer or scientist if your research becomes your hobby,” Kim says. “I wanted to teach this cycle — of happiness, research, and passion — to young people and help PhD students develop like artists or singers.”

In 2015, he packed his bags for MIT, where he accepted a junior faculty position in the Department of Mechanical Engineering. His first impressions upon arriving at the Institute?

“Freedom,” Kim says. “For me, free thinking — to compose music, innovate something totally new — is the most important thing. And the people at MIT are very talented and curious of all the things.”

Since he’s put down roots on campus, he has built up a highly productive research group, focused on fabricating ultra-thin, stackable, high-performance electronic materials and devices, which Kim envisions could be used to build hybrid electronic systems as small as a fingernail and as powerful as a supercomputer. He credits the group’s many innovations to the more than 40 students, postdocs, and research scientists who have contributed to his lab.

“I hope this is where they can learn that research can be an art,” Kim says. “To students especially, I hope they see that, if they enjoy what they do, then they can be whatever they want to be.”

New algorithm unlocks high-resolution insights for computer vision

Mon, 03/18/2024 - 3:10pm

Imagine yourself glancing at a busy street for a few moments, then trying to sketch the scene you saw from memory. Most people could draw the rough positions of the major objects like cars, people, and crosswalks, but almost no one can draw every detail with pixel-perfect accuracy. The same is true for most modern computer vision algorithms: They are fantastic at capturing high-level details of a scene, but they lose fine-grained details as they process information.

Now, MIT researchers have created a system called “FeatUp” that lets algorithms capture all of the high- and low-level details of a scene at the same time — almost like Lasik eye surgery for computer vision.

When computers learn to “see” from looking at images and videos, they build up “ideas” of what's in a scene through something called “features.” To create these features, deep networks and visual foundation models break down images into a grid of tiny squares and process these squares as a group to determine what's going on in a photo. Each tiny square is usually made up of anywhere from 16 to 32 pixels, so the resolution of these algorithms is dramatically smaller than the images they work with. In trying to summarize and understand photos, algorithms lose a ton of pixel clarity. 

The FeatUp algorithm can stop this loss of information and boost the resolution of any deep network without compromising on speed or quality. This allows researchers to quickly and easily improve the resolution of any new or existing algorithm. For example, imagine trying to interpret the predictions of a lung cancer detection algorithm with the goal of localizing the tumor. Applying FeatUp before interpreting the algorithm using a method like class activation maps (CAM) can yield a dramatically more detailed (16-32x) view of where the tumor might be located according to the model. 

FeatUp not only helps practitioners understand their models, but also can improve a panoply of different tasks like object detection, semantic segmentation (assigning labels to pixels in an image with object labels), and depth estimation. It achieves this by providing more accurate, high-resolution features, which are crucial for building vision applications ranging from autonomous driving to medical imaging.

“The essence of all computer vision lies in these deep, intelligent features that emerge from the depths of deep learning architectures. The big challenge of modern algorithms is that they reduce large images to  very small grids of 'smart' features, gaining intelligent insights but losing the finer details,” says Mark Hamilton, an MIT PhD student in electrical engineering and computer science, MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) affiliate, and a co-lead author on a paper about the project. “FeatUp helps enable the best of both worlds: highly intelligent representations with the original image’s resolution. These high-resolution features significantly boost performance across a spectrum of computer vision tasks, from enhancing object detection and improving depth prediction to providing a deeper understanding of your network's decision-making process through high-resolution analysis.” 

Resolution renaissance 

As these large AI models become more and more prevalent, there’s an increasing need to explain what they’re doing, what they’re looking at, and what they’re thinking. 

But how exactly can FeatUp discover these fine-grained details? Curiously, the secret lies in wiggling and jiggling images. 

In particular, FeatUp applies minor adjustments (like moving the image a few pixels to the left or right) and watches how an algorithm responds to these slight movements of the image. This results in hundreds of deep-feature maps that are all slightly different, which can be combined into a single crisp, high-resolution, set of deep features. “We imagine that some high-resolution features exist, and that when we wiggle them and blur them, they will match all of the original, lower-resolution features from the wiggled images. Our goal is to learn how to refine the low-resolution features into high-resolution features using this 'game' that lets us know how well we are doing,” says Hamilton. This methodology is analogous to how algorithms can create a 3D model from multiple 2D images by ensuring that the predicted 3D object matches all of the 2D photos used to create it. In FeatUp’s case, they predict a high-resolution feature map that’s consistent with all of the low-resolution feature maps formed by jittering the original image.

The team notes that standard tools available in PyTorch were insufficient for their needs, and introduced a new type of deep network layer in their quest for a speedy and efficient solution. Their custom layer, a special joint bilateral upsampling operation, was over 100 times more efficient than a naive implementation in PyTorch. The team also showed this new layer could improve a wide variety of different algorithms including semantic segmentation and depth prediction. This layer improved the network’s ability to process and understand high-resolution details, giving any algorithm that used it a substantial performance boost. 

“Another application is something called small object retrieval, where our algorithm allows for precise localization of objects. For example, even in cluttered road scenes algorithms enriched with FeatUp can see tiny objects like traffic cones, reflectors, lights, and potholes where their low-resolution cousins fail. This demonstrates its capability to enhance coarse features into finely detailed signals,” says Stephanie Fu ’22, MNG ’23, a PhD student at the University of California at Berkeley and another co-lead author on the new FeatUp paper. “This is especially critical for time-sensitive tasks, like pinpointing a traffic sign on a cluttered expressway in a driverless car. This can not only improve the accuracy of such tasks by turning broad guesses into exact localizations, but might also make these systems more reliable, interpretable, and trustworthy.”

What next?

Regarding future aspirations, the team emphasizes FeatUp’s potential widespread adoption within the research community and beyond, akin to data augmentation practices. “The goal is to make this method a fundamental tool in deep learning, enriching models to perceive the world in greater detail without the computational inefficiency of traditional high-resolution processing,” says Fu.

“FeatUp represents a wonderful advance towards making visual representations really useful, by producing them at full image resolutions,” says Cornell University computer science professor Noah Snavely, who was not involved in the research. “Learned visual representations have become really good in the last few years, but they are almost always produced at very low resolution — you might put in a nice full-resolution photo, and get back a tiny, postage stamp-sized grid of features. That’s a problem if you want to use those features in applications that produce full-resolution outputs. FeatUp solves this problem in a creative way by combining classic ideas in super-resolution with modern learning approaches, leading to beautiful, high-resolution feature maps.”

“We hope this simple idea can have broad application. It provides high-resolution versions of image analytics that we’d thought before could only be low-resolution,” says senior author William T. Freeman, an MIT professor of electrical engineering and computer science professor and CSAIL member.

Lead authors Fu and Hamilton are accompanied by MIT PhD students Laura Brandt SM ’21 and Axel Feldmann SM ’21, as well as Zhoutong Zhang SM ’21, PhD ’22, all current or former affiliates of MIT CSAIL. Their research is supported, in part, by a National Science Foundation Graduate Research Fellowship, by the National Science Foundation and Office of the Director of National Intelligence, by the U.S. Air Force Research Laboratory, and by the U.S. Air Force Artificial Intelligence Accelerator. The group will present their work in May at the International Conference on Learning Representations.

Five MIT faculty members take on Cancer Grand Challenges

Mon, 03/18/2024 - 10:15am

Cancer Grand Challenges recently announced five winning teams for 2024, which included five researchers from MIT: Michael Birnbaum, Regina Barzilay, Brandon DeKosky, Seychelle Vos, and Ömer Yilmaz. Each team is made up of interdisciplinary cancer researchers from across the globe and will be awarded $25 million over five years. 

Birnbaum, an associate professor in the Department of Biological Engineering, leads Team MATCHMAKERS and is joined by co-investigators Barzilay, the School of Engineering Distinguished Professor for AI and Health in the Department of Electrical Engineering and Computer Science and the AI faculty lead at the MIT Abdul Latif Jameel Clinic for Machine Learning in Health; and DeKosky, Phillip and Susan Ragon Career Development Professor of Chemical Engineering. All three are also affiliates of the Koch Institute for Integrative Cancer Research At MIT.

Team MATCHMAKERS will take advantage of recent advances in artificial intelligence to develop tools for personalized immunotherapies for cancer patients. Cancer immunotherapies, which recruit the patient’s own immune system against the disease, have transformed treatment for some cancers, but not for all types and not for all patients. 

T cells are one target for immunotherapies because of their central role in the immune response. These immune cells use receptors on their surface to recognize protein fragments called antigens on cancer cells. Once T cells attach to cancer antigens, they mark them for destruction by the immune system. However, T cell receptors are exceptionally diverse within one person’s immune system and from person to person, making it difficult to predict how any one cancer patient will respond to an immunotherapy.  

Team MATCHMAKERS will collect data on T cell receptors and the different antigens they target and build computer models to predict antigen recognition by different T cell receptors. The team’s overarching goal is to develop tools for predicting T cell recognition with simple clinical lab tests and designing antigen-specific immunotherapies. “If successful, what we learn on our team could help transform prediction of T cell receptor recognition from something that is only possible in a few sophisticated laboratories in the world, for a few people at a time, into a routine process,” says Birnbaum. 

“The MATCHMAKERS project draws on MIT’s long tradition of developing cutting-edge artificial intelligence tools for the benefit of society,” comments Ryan Schoenfeld, CEO of The Mark Foundation for Cancer Research. “Their approach to optimizing immunotherapy for cancer and many other diseases is exemplary of the type of interdisciplinary research The Mark Foundation prioritizes supporting.” In addition to The Mark Foundation, the MATCHMAKERS team is funded by Cancer Research UK and the U.S. National Cancer Institute.

Vos, the Robert A. Swanson (1969) Career Development Professor of Life Sciences and HHMI Freeman Hrabowksi Scholar in the Department of Biology, will be a co-investigator on Team KOODAC. The KOODAC team will develop new treatments for solid tumors in children, using protein degradation strategies to target previously “undruggable” drivers of cancers. KOODAC is funded by Cancer Research UK, France's Institut National Du Cancer, and KiKa (Children Cancer Free Foundation) through Cancer Grand Challenges. 

As a co-investigator on team PROSPECT, Yilmaz, who is also a Koch Institute affiliate, will help address early-onset colorectal cancers, an emerging global problem among individuals younger than 50 years. The team seeks to elucidate pathways, risk factors, and molecules involved in the disease’s development. Team PROSPECT is supported by Cancer Research UK, the U.S. National Cancer Institute, the Bowelbabe Fund for Cancer Research UK, and France's Institut National Du Cancer through Cancer Grand Challenges.  

Unlocking the quantum future

Mon, 03/18/2024 - 9:55am

Quantum computing is the next frontier for faster and more powerful computing technologies. It has the potential to better optimize routes for shipping and delivery, speed up battery development for electric vehicles, and more accurately predict trends in financial markets. But to unlock the quantum future, scientists and engineers need to solve outstanding technical challenges while continuing to explore new applications.

One place where they’re working towards this future is the MIT Interdisciplinary Quantum Hackathon, or iQuHACK for short (pronounced “i-quack,” like a duck). Each year, a community of quhackers (quantum hackers) gathers at iQuHACK to work on quantum computing projects using real quantum computers and simulators. This year, the hackathon was held both in-person at MIT and online over three days in February.

Quhackers worked in teams to advance the capability of quantum computers and to investigate promising applications. Collectively, they tackled a wide range of projects, such as running a quantum-powered dating service, building an organ donor matching app, and breaking into quantum vaults. While working, quhackers could consult with scientists and engineers in attendance from sponsoring companies. Many sponsors also received feedback and ideas from quhackers to help improve their quantum platforms.

But organizing iQuHACK 2024 was no easy feat. Co-chairs Alessandro Buzzi and Daniela Zaidenberg led a committee of nine members to hold the largest iQuHACK yet. “It wouldn’t have been possible without them,” Buzzi said. The hackathon hosted 260 in-person quhackers and 1,000 remote quhackers, representing 77 countries in total. More than 20 scientists and engineers from sponsoring companies also attended in person as mentors for quhackers.

Each team of quhackers tackled one of 10 challenges posed by the hackathon’s eight major sponsoring companies. Some challenges asked quhackers to improve computing performance, such as by making quantum algorithms faster and more accurate. Other challenges asked quhackers to explore applying quantum computing to other fields, such as finance and machine learning. The sponsors worked with the iQuHACK committee to craft creative challenges with industry relevance and societal impact. “We wanted people to be able to address an interesting challenge [that has] applications in the real world,” says Zaidenberg.

One team of quhackers looked for potential quantum applications and found one close to home: dating. A team member, Liam Kronman, had previously built dating apps but disliked that matching algorithms for normal classical computers “require [an overly] strict setup.” With these classical algorithms, people must be split into two groups — for example, men and women — and matches can only be made between these groups. But with quantum computers, matching algorithms are more flexible and can consider all possible combinations, enabling the inclusion of multiple genders and gender preferences. 

Kronman and his team members leveraged these quantum algorithms to build a quantum-powered dating platform called MITqute (pronounced “meet cute”). To date, the platform has matched at least 240 people from the iQuHACK and MIT undergrad communities. In a follow-up survey, 13 out of 41 respondents reported having talked with their match, with at least two pairs setting up dates. “I really lucked out with this one,” one respondent wrote. 

Another team of quhackers also based their project on quantum matching algorithms but instead leveraged the algorithms’ power for medical care. The team built a mobile app that matches organ donors to patients, earning them the hackathon’s top social impact award. 

But they almost didn’t go through with their project. “At one point, we were considering scrapping the whole thing because we thought we couldn’t implement the algorithm,” says Alma Alex, one of the developers. After talking with their hackathon mentor for advice, though, the team learned that another group was working on a similar type of project — incidentally, the MITqute team. Knowing that others were tackling the same problem inspired them to persevere.

A sense of community also helped to motivate other quhackers. For one of the challenges, quhackers were tasked with hacking into 13 virtual quantum vaults. Teams could see each other’s progress on each vault in real time on a leaderboard, and this knowledge informed their strategies. When the first vault was successfully hacked by a team, progress from many other teams spiked on that vault and slowed down on others, says Daiwei Zhu, a quantum applications scientist at IonQ and one of the challenge’s two architects.

The vault challenge may appear to be just a fun series of puzzles, but the solutions can be used in quantum computers to improve their efficiency and accuracy. To hack into a vault, quhackers had to first figure out its secret key — an unknown quantum state — using a maximum of 20 probing tests. Then, they had to change the key’s state to a target state. These types of characterizations and modifications of quantum states are “fundamental” for quantum computers to work, says Jason Iaconis, a quantum applications engineer at IonQ and the challenge’s other architect. 

But the best way to characterize and modify states is not yet clear. “Some of the [vaults] we [didn’t] even know how to solve ourselves,” Zhu says. At the end of the hackathon, six vaults had at least one team mostly hack into them. (In the quantum world where gray areas exist, it’s possible to partly hack into a vault.)

The community of scientists and engineers formed at iQuHACK persists beyond the weekend, and many members continue to grow the community outside the hackathon. Inspired quhackers have gone on to start their own quantum computing clubs at their universities. A few years ago, a group of undergraduate quhackers from different universities formed a Quantum Coalition that now hosts their own quantum hackathons. “It’s crazy to see how the hackathon itself spreads and how many people start their own initiatives,” co-chair Zaidenberg says. 

The three-day hackathon opened with a keynote from MIT Professor Will Oliver, which included an overview of basic quantum computing concepts, current challenges, and computing technologies. Following that were industry talks and a panel of six industry and academic quantum experts, including MIT Professor Peter Shor, who is known for developing one of the most famous quantum algorithms. The panelists discussed current challenges, future applications, the importance of collaboration, and the need for ample testing.

Later, sponsors held technical workshops where quhackers could learn the nitty-gritty details of programming on specific quantum platforms. Day one closed out with a talk by research scientist Xinghui Yin on the role of quantum technology at LIGO, the Laser Interferometer Gravitational-Wave Observatory that first detected gravitational waves. The next day, the hackathon’s challenges were announced at 10 a.m., and hacking kicked off at the MIT InnovationHQ. In the afternoon, attendees could also tour MIT quantum computing labs.

Hacking continued overnight at the MIT Museum and ended back at MIT iHQ at 10 a.m. on the final day. Quhackers then presented their projects to panels of judges. Afterward, industry speakers gave lightning talks about each of their company’s latest quantum technologies and future directions. The hackathon ended with a closing ceremony, where sponsors announced the awards for each of the 10 challenges. 

The hackathon was captured in a three-part video by Albert Figurt, a resident artist at MIT. Figurt shot and edited the footage in parallel with the hackathon. Each part represented one day of the hackathon and was released on the subsequent day.

Throughout the weekend, quhackers and sponsors consistently praised the hackathon’s execution and atmosphere. “That was amazing … never felt so much better, one of the best hackathons I did from over 30 hackathons I attended,” Abdullah Kazi, a quhacker, wrote on the iQuHACK Slack.

Ultimately, “[we wanted to] help people to meet each other,” co-chair Buzzi says. “The impact [of iQuHACK] is scientific in some way, but it’s very human at the most important level.”

Making the clean energy transition work for everyone

Fri, 03/15/2024 - 5:00pm

The clean energy transition is already underway, but how do we make sure it happens in a manner that is affordable, sustainable, and fair for everyone?

That was the overarching question at this year’s MIT Energy Conference, which took place March 11 and 12 in Boston and was titled “Short and Long: A Balanced Approach to the Energy Transition.”

Each year, the student-run conference brings together leaders in the energy sector to discuss the progress and challenges they see in their work toward a greener future. Participants come from research, industry, government, academia, and the investment community to network and exchange ideas over two whirlwind days of keynote talks, fireside chats, and panel discussions.

Several participants noted that clean energy technologies are already cost-competitive with fossil fuels, but changing the way the world works requires more than just technology.

“None of this is easy, but I think developing innovative new technologies is really easy compared to the things we’re talking about here, which is how to blend social justice, soft engineering, and systems thinking that puts people first,” Daniel Kammen, a distinguished professor of energy at the University of California at Berkeley, said in a keynote talk. “While clean energy has a long way to go, it is more than ready to transition us from fossil fuels.”

The event also featured a keynote discussion between MIT President Sally Kornbluth and MIT’s Kyocera Professor of Ceramics Yet-Ming Chiang, in which Kornbluth discussed her first year at MIT as well as a recently announced, campus-wide effort to solve critical climate problems known as the Climate Project at MIT.

“The reason I wanted to come to MIT was I saw that MIT has the potential to solve the world’s biggest problems, and first among those for me was the climate crisis,” Kornbluth said. “I’m excited about where we are, I’m excited about the enthusiasm of the community, and I think we’ll be able to make really impactful discoveries through this project.”

Fostering new technologies

Several panels convened experts in new or emerging technology fields to discuss what it will take for their solutions to contribute to deep decarbonization.

“The fun thing and challenging thing about first-of-a-kind technologies is they’re all kind of different,” said Jonah Wagner, principal assistant director for industrial innovation and clean energy in the U.S. Office of Science and Technology Policy. “You can map their growth against specific challenges you expect to see, but every single technology is going to face their own challenges, and every single one will have to defy an engineering barrier to get off the ground.”

Among the emerging technologies discussed was next-generation geothermal energy, which uses new techniques to extract heat from the Earth’s crust in new places.

A promising aspect of the technology is that it can leverage existing infrastructure and expertise from the oil and gas industry. Many newly developed techniques for geothermal production, for instance, use the same drills and rigs as those used for hydraulic fracturing.

“The fact that we have a robust ecosystem of oil and gas labor and technology in the U.S. makes innovation in geothermal much more accessible compared to some of the challenges we’re seeing in nuclear or direct-air capture, where some of the supply chains are disaggregated around the world,” said Gabrial Malek, chief of staff at the geothermal company Fervo Energy.

Another technology generating excitement — if not net energy quite yet — is fusion, the process of combining, or fusing, light atoms together to form heavier ones for a net energy gain, in the same process that powers the sun. MIT spinout Commonwealth Fusion Systems (CFS) has already validated many aspects of its approach for achieving fusion power, and the company’s unique partnership with MIT was discussed in a panel on the industry’s progress.

“We’re standing on the shoulders of decades of research from the scientific community, and we want to maintain those ties even as we continue developing our technology,” CFS Chief Science Officer Brandon Sorbom PhD ’17 said, noting that CFS is one of the largest company sponsors of research at MIT and collaborates with institutions around the world. “Engaging with the community is a really valuable lever to get new ideas and to sanity check our own ideas.”

Sorbom said that as CFS advances fusion energy, the company is thinking about how it can replicate its processes to lower costs and maximize the technology’s impact around the planet.

“For fusion to work, it has to work for everyone,” Sorbom said. “I think the affordability piece is really important. We can’t just build this technological jewel that only one class of nations can afford. It has to be a technology that can be deployed throughout the entire world.”

The event also gave students — many from MIT — a chance to learn more about careers in energy and featured a startup showcase, in which dozens of companies displayed their energy and sustainability solutions.

“More than 700 people are here from every corner of the energy industry, so there are so many folks to connect with and help me push my vision into reality,” says GreenLIB CEO Fred Rostami, whose company recycles lithium-ion batteries. “The good thing about the energy transition is that a lot of these technologies and industries overlap, so I think we can enable this transition by working together at events like this.”

A focused climate strategy

Kornbluth noted that when she came to MIT, a large percentage of students and faculty were already working on climate-related technologies. With the Climate Project at MIT, she wanted to help ensure the whole of those efforts is greater than the sum of its parts.

The project is organized around six distinct missions, including decarbonizing energy and industry, empowering frontline communities, and building healthy, resilient cities. Kornbluth says the mission areas will help MIT community members collaborate around multidisciplinary challenges. Her team, which includes a committee of faculty advisors, has begun to search for the leads of each mission area, and Kornbluth said she is planning to appoint a vice president for climate at the Institute.

“I want someone who has the purview of the whole Institute and will report directly to me to help make sure this project stays on track,” Kornbluth explained.

In his conversation about the initiative with Kornbluth, Yet-Ming Chiang said projects will be funded based on their potential to reduce emissions and make the planet more sustainable at scale.

“Projects should be very high risk, with very high impact,” Chiang explained. “They should have a chance to prove themselves, and those efforts should not be limited by resources, only by time.”

In discussing her vision of the climate project, Kornbluth alluded to the “short and long” theme of the conference.

“It’s about balancing research and commercialization,” Kornbluth said. “The climate project has a very variable timeframe, and I think universities are the sector that can think about the things that might be 30 years out. We have to think about the incentives across the entire innovation pipeline and how we can keep an eye on the long term while making sure the short-term things get out rapidly.”

3 Questions: What you need to know about audio deepfakes

Fri, 03/15/2024 - 4:50pm

Audio deepfakes have had a recent bout of bad press after an artificial intelligence-generated robocall purporting to be the voice of Joe Biden hit up New Hampshire residents, urging them not to cast ballots. Meanwhile, spear-phishers — phishing campaigns that target a specific person or group, especially using information known to be of interest to the target — go fishing for money, and actors aim to preserve their audio likeness.

What receives less press, however, are some of the uses of audio deepfakes that could actually benefit society. In this Q&A prepared for MIT News, postdoc Nauman Dawalatabad addresses concerns as well as potential upsides of the emerging tech. A fuller version of this interview can be seen at the video below.

Q: What ethical considerations justify the concealment of the source speaker's identity in audio deepfakes, especially when this technology is used for creating innovative content?

A: The inquiry into why research is important in obscuring the identity of the source speaker, despite a large primary use of generative models for audio creation in entertainment, for example, does raise ethical considerations. Speech does not contain the information only about “who you are?” (identity) or “what you are speaking?” (content); it encapsulates a myriad of sensitive information including age, gender, accent, current health, and even cues about the upcoming future health conditions. For instance, our recent research paper on “Detecting Dementia from Long Neuropsychological Interviews” demonstrates the feasibility of detecting dementia from speech with considerably high accuracy. Moreover, there are multiple models that can detect gender, accent, age, and other information from speech with very high accuracy. There is a need for advancements in technology that safeguard against the inadvertent disclosure of such private data. The endeavor to anonymize the source speaker's identity is not merely a technical challenge but a moral obligation to preserve individual privacy in the digital age.

Q: How can we effectively maneuver through the challenges posed by audio deepfakes in spear-phishing attacks, taking into account the associated risks, the development of countermeasures, and the advancement of detection techniques?

A: The deployment of audio deepfakes in spear-phishing attacks introduces multiple risks, including the propagation of misinformation and fake news, identity theft, privacy infringements, and the malicious alteration of content. The recent circulation of deceptive robocalls in Massachusetts exemplifies the detrimental impact of such technology. We also recently spoke with the spoke with The Boston Globe about this technology, and how easy and inexpensive it is to generate such deepfake audios.

Anyone without a significant technical background can easily generate such audio, with multiple available tools online. Such fake news from deepfake generators can disturb financial markets and even electoral outcomes. The theft of one's voice to access voice-operated bank accounts and the unauthorized utilization of one's vocal identity for financial gain are reminders of the urgent need for robust countermeasures. Further risks may include privacy violation, where an attacker can utilize the victim’s audio without their permission or consent. Further, attackers can also alter the content of the original audio, which can have a serious impact.

Two primary and prominent directions have emerged in designing systems to detect fake audio: artifact detection and liveness detection. When audio is generated by a generative model, the model introduces some artifact in the generated signal. Researchers design algorithms/models to detect these artifacts. However, there are some challenges with this approach due to increasing sophistication of audio deepfake generators. In the future, we may also see models with very small or almost no artifacts. Liveness detection, on the other hand, leverages the inherent qualities of natural speech, such as breathing patterns, intonations, or rhythms, which are challenging for AI models to replicate accurately. Some companies like Pindrop are developing such solutions for detecting audio fakes. 

Additionally, strategies like audio watermarking serve as proactive defenses, embedding encrypted identifiers within the original audio to trace its origin and deter tampering. Despite other potential vulnerabilities, such as the risk of replay attacks, ongoing research and development in this arena offer promising solutions to mitigate the threats posed by audio deepfakes.

Q: Despite their potential for misuse, what are some positive aspects and benefits of audio deepfake technology? How do you imagine the future relationship between AI and our experiences of audio perception will evolve?

A: Contrary to the predominant focus on the nefarious applications of audio deepfakes, the technology harbors immense potential for positive impact across various sectors. Beyond the realm of creativity, where voice conversion technologies enable unprecedented flexibility in entertainment and media, audio deepfakes hold transformative promise in health care and education sectors. My current ongoing work in the anonymization of patient and doctor voices in cognitive health-care interviews, for instance, facilitates the sharing of crucial medical data for research globally while ensuring privacy. Sharing this data among researchers fosters development in the areas of cognitive health care. The application of this technology in voice restoration represents a hope for individuals with speech impairments, for example, for ALS or dysarthric speech, enhancing communication abilities and quality of life.

I am very positive about the future impact of audio generative AI models. The future interplay between AI and audio perception is poised for groundbreaking advancements, particularly through the lens of psychoacoustics — the study of how humans perceive sounds. Innovations in augmented and virtual reality, exemplified by devices like the Apple Vision Pro and others, are pushing the boundaries of audio experiences towards unparalleled realism. Recently we have seen an exponential increase in the number of sophisticated models coming up almost every month. This rapid pace of research and development in this field promises not only to refine these technologies but also to expand their applications in ways that profoundly benefit society. Despite the inherent risks, the potential for audio generative AI models to revolutionize health care, entertainment, education, and beyond is a testament to the positive trajectory of this research field.

Study finds lands used for grazing can worsen or help climate change

Fri, 03/15/2024 - 6:00am

When it comes to global climate change, livestock grazing can be either a blessing or a curse, according to a new study, which offers clues on how to tell the difference.

If managed properly, the study shows, grazing can actually increase the amount of carbon from the air that gets stored in the ground and sequestered for the long run. But if there is too much grazing, soil erosion can result, and the net effect is to cause more carbon losses, so that the land becomes a net carbon source, instead of a carbon sink. And the study found that the latter is far more common around the world today.

The new work, published today in the journal Nature Climate Change, provides ways to determine the tipping point between the two, for grazing lands in a given climate zone and soil type. It also provides an estimate of the total amount of carbon that has been lost over past decades due to livestock grazing, and how much could be removed from the atmosphere if grazing optimization management implemented. The study was carried out by Cesar Terrer, an assistant professor of civil and environmental engineering at MIT; Shuai Ren, a PhD student at the Chinese Academy of Sciences whose thesis is co-supervised by Terrer; and four others.

“This has been a matter of debate in the scientific literature for a long time,” Terrer says. “In general experiments, grazing decreases soil carbon stocks, but surprisingly, sometimes grazing increases soil carbon stocks, which is why it’s been puzzling.”

What happens, he explains, is that “grazing could stimulate vegetation growth through easing resource constraints such as light and nutrients, thereby increasing root carbon inputs to soils, where carbon can stay there for centuries or millennia.”

But that only works up to a certain point, the team found after a careful analysis of 1,473 soil carbon observations from different grazing studies from many locations around the world. “When you cross a threshold in grazing intensity, or the amount of animals grazing there, that is when you start to see sort of a tipping point — a strong decrease in the amount of carbon in the soil,” Terrer explains.

That loss is thought to be primarily from increased soil erosion on the denuded land. And with that erosion, Terrer says, “basically you lose a lot of the carbon that you have been locking in for centuries.”

The various studies the team compiled, although they differed somewhat, essentially used similar methodology, which is to fence off a portion of land so that livestock can’t access it, and then after some time take soil samples from within the enclosure area, and from comparable nearby areas that have been grazed, and compare the content of carbon compounds.

“Along with the data on soil carbon for the control and grazed plots,” he says, “we also collected a bunch of other information, such as the mean annual temperature of the site, mean annual precipitation, plant biomass, and properties of the soil, like pH and nitrogen content. And then, of course, we estimate the grazing intensity — aboveground biomass consumed, because that turns out to be the key parameter.”  

With artificial intelligence models, the authors quantified the importance of each of these parameters, those drivers of intensity — temperature, precipitation, soil properties — in modulating the sign (positive or negative) and magnitude of the impact of grazing on soil carbon stocks. “Interestingly, we found soil carbon stocks increase and then decrease with grazing intensity, rather than the expected linear response,” says Ren.

Having developed the model through AI methods and validated it, including by comparing its predictions with those based on underlying physical principles, they can then apply the model to estimating both past and future effects. “In this case,” Terrer says, “we use the model to quantify the historical loses in soil carbon stocks from grazing. And we found that 46 petagrams [billion metric tons] of soil carbon, down to a depth of one meter, have been lost in the last few decades due to grazing.”

By way of comparison, the total amount of greenhouse gas emissions per year from all fossil fuels is about 10 petagrams, so the loss from grazing equals more than four years’ worth of all the world’s fossil emissions combined.

What they found was “an overall decline in soil carbon stocks, but with a lot of variability.” Terrer says. The analysis showed that the interplay between grazing intensity and environmental conditions such as temperature could explain the variability, with higher grazing intensity and hotter climates resulting in greater carbon loss. “This means that policy-makers should take into account local abiotic and biotic factors to manage rangelands efficiently,” Ren notes. “By ignoring such complex interactions, we found that using IPCC [Intergovernmental Panel on Climate Change] guidelines would underestimate grazing-induced soil carbon loss by a factor of three globally.”

Using an approach that incorporates local environmental conditions, the team produced global, high-resolution maps of optimal grazing intensity and the threshold of intensity at which carbon starts to decrease very rapidly. These maps are expected to serve as important benchmarks for evaluating existing grazing practices and provide guidance to local farmers on how to effectively manage their grazing lands.

Then, using that map, the team estimated how much carbon could be captured if all grazing lands were limited to their optimum grazing intensity. Currently, the authors found, about 20 percent of all pasturelands have crossed the thresholds, leading to severe carbon losses. However, they found that under the optimal levels, global grazing lands would sequester 63 petagrams of carbon. “It is amazing,” Ren says. “This value is roughly equivalent to a 30-year carbon accumulation from global natural forest regrowth.”

That would be no simple task, of course. To achieve optimal levels, the team found that approximately 75 percent of all grazing areas need to reduce grazing intensity. Overall, if the world seriously reduces the amount of grazing, “you have to reduce the amount of meat that’s available for people,” Terrer says.

“Another option is to move cattle around,” he says, “from areas that are more severely affected by grazing intensity, to areas that are less affected. Those rotations have been suggested as an opportunity to avoid the more drastic declines in carbon stocks without necessarily reducing the availability of meat.”

This study didn’t delve into these social and economic implications, Terrer says. “Our role is to just point out what would be the opportunity here. It shows that shifts in diets can be a powerful way to mitigate climate change.”

“This is a rigorous and careful analysis that provides our best look to date at soil carbon changes due to livestock grazing practiced worldwide,” say Ben Bond-Lamberty, a terrestrial ecosystem research scientist at Pacific Northwest National Laboratory, who was not associated with this work. “The authors’ analysis gives us a unique estimate of soil carbon losses due to grazing and, intriguingly, where and how the process might be reversed.”

He adds: “One intriguing aspect to this work is the discrepancies between its results and the guidelines currently used by the IPCC — guidelines that affect countries’ commitments, carbon-market pricing, and policies.” However, he says, “As the authors note, the amount of carbon historically grazed soils might be able to take up is small relative to ongoing human emissions. But every little bit helps!”

“Improved management of working lands can be a powerful tool to combat climate change,” says Jonathan Sanderman, carbon program director of the Woodwell Climate Research Center in Falmouth, Massachusetts, who was not associated with this work. He adds, “This work demonstrates that while, historically, grazing has been a large contributor to climate change, there is significant potential to decrease the climate impact of livestock by optimizing grazing intensity to rebuild lost soil carbon.”

Terrer states that for now, “we have started a new study, to evaluate the consequences of shifts in diets for carbon stocks. I think that’s the million-dollar question: How much carbon could you sequester, compared to business as usual, if diets shift to more vegan or vegetarian?” The answers will not be simple, because a shift to more vegetable-based diets would require more cropland, which can also have different environmental impacts. Pastures take more land than crops, but produce different kinds of emissions. “What’s the overall impact for climate change? That is the question we’re interested in,” he says.

The research team included Juan Li, Yingfao Cao, Sheshan Yang, and Dan Liu, all with the  Chinese Academy of Sciences. The work was supported by the Second Tibetan Plateau Scientific Expedition and Research Program, and the Science and Technology Major Project of Tibetan Autonomous Region of China.

Envisioning a time when people age without fear of dementia

Fri, 03/15/2024 - 12:00am

The mathematician and computer scientist Richard Hamming once gave a talk about doing great research. “He who works with the door open gets all kinds of interruptions, but he also occasionally gets clues as to what the world is and what might be important,” Hamming said, emphasizing the importance of open-mindedness and scientific development.

William Li came across this quote as a high school student seeking to dedicate himself to research but unsure how to begin. “I think that science is kind of an opaque area to break into. It’s hard to know what you’re supposed to be doing from time to time,” Li explains.

A double-major in physics and computer science, Li has taken this advice to heart. Keeping his “office door” open has led him to a variety of research projects, from neuroimaging to genomics, that shaped his long-term goal: to become a physician-scientist who moves the needle on Alzheimer’s disease.

Li’s interest in working with patients in a clinical setting was spurred by his grandfather, who was a doctor. In high school, he began volunteering in retirement homes and at the Byrd Alzheimer’s Center and Research Institute at the University of South Florida. Through this work, Li witnessed the devastating effects Alzheimer’s disease has on both those diagnosed and their loved ones.

But that isn’t the only thing about Alzheimer’s that has grabbed his interest. With no cure available, and relatively little known about its cause, the disease is also a compelling scientific problem. “Beyond its human impact, Alzheimer’s represents a frontier of our understanding of human disease,” Li says.

Starting in the fall, Li will begin an MD/PhD program “for the better part of the coming decade.” Following that, he hopes to secure a residency in radiology or neurology, and then to teach and do research while simultaneously practicing medicine. His ultimate goal is a big one — to help develop a cure for Alzheimer’s.

Pursing knowledge

Research has been the highlight of Li’s career at MIT. He says, “To me personally, research means being able to contribute to a body of knowledge built upon by generations of minds in the past. I see modern science and technology as a pinnacle of human achievement, and it’s a dream come true to be able to add to this work.”

In a normal week during the academic semester, Li can spend up to 15 hours in the lab. His research projects have addressed very different topics, but both have guided him toward his current goals.

In the Soljačić and Johnson groups in the Research Laboratory of Electronics, Li he has worked in nanophotonics, a field concerned with controlling light by designing structures the size of a wavelength, for optical and X-ray images, among other applications. 

Li has worked on making X-ray imaging safer and more effective for medical screenings. He also focuses on using computational methods to design nanophotonic device elements for higher-resolution imaging. “Imaging technologies in the future will have pretty enormous applications both for understanding disease and for being able to catch diseases early through diagnosis,” he says.

In his sophomore year, Li began working at the lab of Professor Manolis Kellis at the Broad Institute of MIT and Harvard, using computational tools to study genetic variation among Alzheimer’s patients and how this relates to the disease itself. In this way, the disease can be broken down into subtypes, explains Li, which will make it easier to understand and treat. Last summer, Li won a SuperUROP Outstanding Research Award for this project.

Forging connections

When Li first joined the Kellis lab, the field of genomics seemed vast and overwhelming. To combat this, he started an academic journal club. In the club, Li and his peers would read research papers together and discuss them. In the fashion of a traditional journal club, one person would present at each meeting. Club participants encouraged each other to focus on any research they found exciting, ranging over the past century. As the club has continued, members have started to present their own research to the group as well. “It’s fun seeing what my friends are interested in,” Li says.

Li also served as the collegiate relations co-chair of MIT’s Pre-Medical Society. Here he was responsible for organizing an annual meeting between all pre-med students of the greater Boston area. This mixer was held for pre-med students to other local students and learn from pre-medical advisors and alumni of various Boston schools.

Among the several communities Li is a part of at MIT, his dormitory holds a special place in his heart. Next House, MIT’s largest dorm building, is the place Li has called home since his junior year. Since moving in, he has immersed himself in the living community by assuming roles in several activities hosted by the dorm, such as Thanksgiving dinner.

“I’m very happy to be part of the Next House community. It’s a pretty fantastic place, and I would say that my quality of social life has increased a lot since moving here,” he states.

Along with large events, Li also appreciates the weekly traditions he has created with his Next House friends. Each Sunday, for example, Li joins members of his dorm wing for a 15-minute workout. He says he enjoys exercising in the group setting and frequently attends the gym with his friends, too.

After some downtime on the weekends, Li heads back to the lab and his quest to better understand the brain and how it can be ravaged by dementia. As he continues on his path toward becoming a researcher and physician, he envisions a world where people can age without fear of illness.

2024 MacVicar Faculty Fellows named

Fri, 03/15/2024 - 12:00am

Four outstanding undergraduate teachers and mentors have been named MacVicar Faculty Fellows: professor of electrical engineering and computer science (EECS) Karl Berggren, professor of political science Andrea Campbell, associate professor of music Emily Richmond Pollock, and professor of EECS Vinod Vaikuntanathan.

For more than 30 years, the MacVicar Faculty Fellows Program has recognized exemplary and sustained contributions to undergraduate education at MIT. The program is named in honor of Margaret MacVicar, MIT’s first dean for undergraduate education and founder of the Undergraduate Research Opportunities Program (UROP).

New fellows are chosen each year through a highly competitive nomination process. They receive an annual stipend and are appointed to a 10-year term. Nominations, including letters of support from colleagues, students, and alumni, are reviewed by an advisory committee led by vice chancellor Ian Waitz with final selections made by provost Cynthia Barnhart.

Role models both in and out of the classroom, Berggren, Campbell, Pollock, and Vaikuntanathan join an elite academy of scholars from across the Institute who are committed to curricular innovation; exceptional teaching; collaboration with colleagues; and supporting students through mentorship, leadership, and advising.

Karl Berggren

“It is a great honor to have been selected for this fellowship. It has particularly made me remember the years of dedicated mentoring and support that I’ve received from my colleagues,” says Karl Berggren. “I’ve also learned a great deal over this period from our students by way of their efforts and thoughtful feedback. MIT continuously strives for excellence in undergraduate education, and I feel very lucky to have been part of that effort.”

Karl Berggren is the Joseph F. and Nancy P. Keithley Professor in the Department of EECS. He received his PhD from Harvard University and his BA in physics from Harvard College. Berggren joined MIT in 1996 as a staff member at Lincoln Laboratory before becoming an assistant professor in 2003. He regularly teaches undergraduate EECS offerings including 6.2000, formerly 6.002 (Electrical Circuits), and 6.3400, formerly 6.02 (Introduction to EECS via Communication Networks).

Sahil Pontula ’23 writes, “Professor Berggren turned 6.002 from a mere course requirement into a truly memorable experience that shaped my current research interests and provided a unique perspective … He is devoted not just to educating the next generation of engineers, but also to imbuing in them interdisciplinary problem-solving perspectives that push the frontiers of science forward.”

MacVicar Fellow and professor of EECS Jeffrey Lang notes, “His lectures are polished, presented with humor, and well-appreciated by his students.” Senior Tiffany Louie adds: “He connects with us, inspires us, and welcomes us to ask questions in class and in the greater electrical engineering field.”

Berggren is also deeply invested in the art and science of teaching. Tomás Palacios, professor of EECS, says, “Professor Berggren is genuinely interested in continuously improving the educational experience of our students. He approaches this in the same methodological and quantitative way we typically approach research. He is well-versed in the most modern theories about learning and he is always happy to share … relevant books and papers on the subject.”

Lang agrees, noting that Berggren “reads articles and books that examine and discuss how learning occurs so that he can become a more effective teacher.” He goes on to recall a conversation in which Berggren explained a new form of homework grading. Instead of reducing grades for errors that did not render an obviously flawed result, he helps students extract key takeaways from their assignments and come to correct solutions on their own. Lang notes that a key benefit of this approach is that it allows graders to “work much more quickly yet carefully” and “provides them more time to spend on giving helpful feedback.”

Adding to his long list of contributions, Berggren also oversees the EECS teaching labs. Since assuming this role, he has pursued changes to ensure that students feel comfortable and confident using the space for both coursework and outside projects, developing their critical thinking and hands-on skills.

Faculty head and professor of electrical engineering Joel Voldman applauds Berggren’s efforts: “Since [he] has taken over, the labs are now a place for projects of all sorts, with students being trained on various processes, parts being easy to obtain, equipment readily available … His fundamental mantra has been to make a space that serves students, meets them where they’re at, and helps them get to where they want to go.”

Andrea Campbell

Andrea Campbell received her BA in social studies from Harvard University and her MA and PhD in political science from the University of California at Berkeley. She joined MIT’s Department of Political Science in 2005 and is currently the Arthur and Ruth Sloan Professor of Political Science and director of undergraduate studies.

Professor Campbell regularly teaches classes 17.30 (Making Public Policy), 17.315 (Health Policy), and 17.317 (U.S. Social Policy). Her research examines the relationships between public policies, public opinion, and political behavior.

A unique aspect of Campbell’s teaching style is the personal approach she brings. In 17.315, Campbell shared her own experiences following a tragic accident in her family, which highlighted the real-life challenges that many face navigating America’s health care system.

According to David Singer, department head and the Raphael Dorman-Helen Starbuck Professor of Political Science, Campbell “weaves personal experience into her teaching in powerful ways … Her openness about her experience permits students to share their own … thereby strengthening their own engagement with the material.”

Singer goes on to say, “In all of her classes, [she] encourages students to examine policymaking not as a technocratic exercise, or an exercise of optimization, but rather as a process infused with politics. In steering her pedagogy in this way, she shows her students how to understand the identity and interests of different groups in society, where their relative power emanates from, and how the rules and institutions of the U.S. political system shape policy outcomes on critical issues like LGBTQ rights, gun control, military intervention, and health care.”

Students say her classes are incredibly impactful, lingering with them for years to come. Her former teaching assistant, now Harvard professor, Justin de Benedictis-Kessner wrote, “Andrea’s talents have been an enormous asset … I have seen how many of her former undergraduate students have gone on to successful careers adjacent to her field of public policy in large part because of her inspiration.” Echoing this sentiment, Julia H. Ginder ’19 writes, “her lessons and mentorship have impacted my day-to-day life and career trajectory even five years after graduation.”

Campbell’s work set the stage for wide-ranging improvements to the Course 17 curriculum and under her leadership, public policy has become the most popular minor in the department. Singer writes, “She ensures that required classes in political institutions, economics, and substantive policy areas are regularly taught, and she proselytizes … to students about the importance of understanding policymaking, especially to [those] in engineering and sciences who might otherwise overlook this critically important domain.”

Campbell is heavily involved with undergraduate advising at the department, school, and Institute levels. She is a popular sponsor of UROPs and attracts many undergraduate researchers each year. Campbell is also co-chair of the Gender Equity Committee in the School of Humanities, Arts, and Social Sciences (SHASS) and the Subcommittee on the Communication Requirement (SOCR).

“It is clear that Andrea takes undergraduate teaching extraordinarily seriously, not just when designing her own classes, but when leading the undergraduate program in our department,” says Adam Berinsky, the Mitsui Professor of Political Science.

Beyond her many pedagogical and curricular accomplishments, Singer notes: “Andrea’s students consistently tout her extraordinary degree of personal engagement. She takes the time to get to know students, to mentor them outside the classroom, and to keep them energized in the classroom. Many express gratitude for Andrea’s willingness to go the extra mile — by staying late after class, holding extra office hours, and even inviting students to her home for Thanksgiving dinner.”

On receiving this award Campbell writes, “I am so grateful to my colleagues and students for taking the time to nominate me and so honored to be selected. Teaching and mentoring MIT students is such a joy. I am well aware that some students come through my door just to fulfill a requirement. Others come with genuine enthusiasm and interest. Either way, I love watching them discover how fascinating political science is and how relevant politics and policymaking are for their lives and their futures.”

Emily Richmond Pollock

“I am truly thrilled to become a MacVicar Faculty Fellow. Working with the undergraduates at MIT is such a gift in itself. When I teach, I can only strive to match the students’ creativity and commitment with my own,” says Emily Richmond Pollock.

Pollock joined MIT’s faculty in 2012. She received her BA in music from Harvard University in 2006 and her MA and PhD in music history and literature from the University of California at Berkeley in 2008 and 2012. She was awarded MIT’s Arthur C. Smith Award for meaningful contributions and devotion to undergraduate student life and learning in 2019 and the James A. and Ruth Levitan Teaching Award from the SHASS in 2020. She currently serves on the SOCR, the Subcommittee on the HASS requirements, and is the inaugural undergraduate chair in the SHASS.

Pollock is a dedicated mentor and advisor and testimonials highlight her commitment to student well-being and intellectual development. “Professor Emily Richmond Pollock is a kind, intentional, and dedicated teacher and advisor,” says senior Katherine Reisig. “By fostering such a welcoming community, she helps the MIT music department be a better place. It is clear … [she] cares deeply about her students, not only that we are doing well academically, but also that we are succeeding in life and doing well mentally.”

MacVicar Faculty Fellow and associate professor of literature Marah Gubar agrees: “Emily has long served as a role model for how to support the ‘whole student’ in ways that build community, right wrongs, and infuse more humanity and beauty into our campus.”

MIT colleagues and students praise Pollock’s extensive contributions to curriculum development at the introductory and advanced levels. She regularly teaches class 21M.011 (Introduction to Western Music) and courses on opera, symphonic repertoire, and the advanced seminar for music majors. Her lectures provide lively learning experiences in which her students are encouraged to think critically about music and culture, dive into unfamiliar operas with curiosity, and compare dramatic elements across time periods.

“I came away from 21M.011 not only with a better understanding of Western music, but with new curiosities and questions about music’s role in the world. Professor Pollock’s teaching made me want to learn more  it encouraged lifelong discovery, curiosity, and education,” Reisig says.

Associate professor of music and MacVicar Faculty Fellow Patricia Tang writes, “Professor Pollock continues to grow as a leader in pedagogical innovation, transforming the music history curriculum and being a true inspiration to her colleagues in her devotion to her students. Though these subjects existed in the course catalog before Pollock’s arrival, in all cases she has radically transformed them, infusing new energy and excitement into the curriculum.”

Pollock also champions issues of diversity, equity, and inclusion in the arts and is dedicated to making classical music and opera more accessible while maintaining the intellectual prestige applauded by students. She encourages students to embrace lesser-known works and step outside their comfort zone, even taking students to the opera herself. She has a “strong interest in anti-racist pedagogies and decolonizing music curriculum … [her] pedagogical innovations are numerous,” Tang observes.

About her impact as an advisor, Tang notes: “Professor Pollock genuinely loves getting to know her students … it is really her ‘superpower.’ It is her mission to make sure [they] are not just surviving but thriving in their first year.”

Senior Hana Ro agrees: “Under her guidance, my academic journey has been transformed, and I have gained not only a profound understanding of [this] subject matter but also a sense of belonging and encouragement that has been invaluable during my time at MIT.”

Furthermore, Pollock ensures that students never feel isolated or alone. Graduate student Frederick Ajisafe says, “If she knew that a cohort was taking a demanding class, she would check in with us … In all cases, Emily emphasized her belief in our ability to succeed and her willingness to help us get there.”

Vinod Vaikuntanathan

Vinod Vaikuntanathan is a professor in the Department of EECS. He received his bachelor’s degree in computer science from the Indian Institute of Technology Madras in 2003 and his SM and PhD degrees in computer science from MIT in 2005 and 2009. Vaikuntanathan joined the faculty in 2013 and in recognition of his contributions to teaching and service to students, he received the Harold E. Edgerton Faculty Achievement Award in 2017 and the Ruth and Joel Spira Award for Distinguished Teaching in 2016.

Vaikuntanathan has taught all three EECS undergraduate theoretical computer science subjects including 6.1210, formerly 6.006 (Introduction to Algorithms); 6.1200, formerly 6.042 (Mathematics for Computer Science); and 6.1220, formerly 6.046 (Design and Analysis of Algorithms).

Students say his classes are challenging, yet approachable and inclusive. Helen Propson ’24 writes, “He excels at making complex subjects like cryptography accessible and captivating for all students, creating an atmosphere where every student’s input is highly regarded. He embraces questions and leaves students feeling inspired and motivated to tackle challenging problems, fostering a sense of confidence and a belief in their own abilities.” She goes on to say, “He often describes intricate concepts as ‘magical,’ and his enthusiasm is contagious, making the material come alive in the classroom.”

MIT alumna Anne Kim concurs: “His teaching style is characterized by its clarity, enthusiasm, and a genuine passion for the subject matter. In his classes, he managed to distill complex algorithms into digestible concepts, making the material accessible to students with varying levels of expertise.”

Vaikuntanathan has also made significant contributions to the EECS curriculum. In spring 2022, he, along with fellow professors in the department, led an effort to improve 6.046. According to professor of EECS and MacVicar Fellow Srini Devadas, “designing a new lecture for 6.046 is not easy. Each new lecture is, typically, days of prep work, including preparing to … give the lecture itself and writing and testing problem set questions, quiz questions, and quiz practice questions. Vinod did all this with skill, aplomb, and enthusiasm. His contributions have had a permanent and beneficial effect on 6.046.”

Widely known for his work in cryptography, including homomorphic encryption and computational complexity, Vaikuntanathan became the lecturer-in-charge of the department’s first theoretical cryptography offering, 6.875. In addition, as the fields of quantum and post-quantum cryptography have grown, “Vinod has added relevant modules to the syllabus, taking the place of topics which had grown obsolete,” Devadas remarks. “Some professors might see teaching the same class multiple times as a chance to save themselves work by reusing the same materials. Vinod sees teaching 6.875 every fall as an opportunity to keep improving the class.”

Vinod Vaikuntanathan is also a devoted mentor and advisor, assisting with first-year UROPs and encouraging students to take advantage of his “open-door” policy. Kim writes that Professor Vaikuntanathan is benefiting her career still as “his mentorship ... extends beyond the classroom through his research” and that he “has mentored and advised dozens of [her] friends in the cryptography space.”

“His encouraging demeanor sets a remarkable example of the kind of teacher every student hopes to encounter during their academic career,” says Propson.

On becoming a MacVicar Faculty Fellow Vaikuntanathan writes, “It is humbling to be in the company of such amazing teachers and mentors, many of whom I have come to think of as my role models. Many thanks to my colleagues and my students for considering me worthy of this honor.”

Researchers help robots navigate efficiently in uncertain environments

Thu, 03/14/2024 - 12:00am

If a robot traveling to a destination has just two possible paths, it needs only to compare the routes’ travel time and probability of success. But if the robot is traversing a complex environment with many possible paths, choosing the best route amid so much uncertainty can quickly become an intractable problem.

MIT researchers developed a method that could help this robot efficiently reason about the best routes to its destination. They created an algorithm for constructing roadmaps of an uncertain environment that balances the tradeoff between roadmap quality and computational efficiency, enabling the robot to quickly find a traversable route that minimizes travel time.

The algorithm starts with paths that are certain to be safe and automatically finds shortcuts the robot could take to reduce the overall travel time. In simulated experiments, the researchers found that their algorithm can achieve a better balance between planning performance and efficiency in comparison to other baselines, which prioritize one or the other.

This algorithm could have applications in areas like exploration, perhaps by helping a robot plan the best way to travel to the edge of a distant crater across the uneven surface of Mars. It could also aid a search-and-rescue drone in finding the quickest route to someone stranded on a remote mountainside.

“It is unrealistic, especially in very large outdoor environments, that you would know exactly where you can and can’t traverse. But if we have just a little bit of information about our environment, we can use that to build a high-quality roadmap,” says Yasmin Veys, an electrical engineering and computer science (EECS) graduate student and lead author of a paper on this technique.

Veys wrote the paper with Martina Stadler Kurtz, a graduate student in the MIT Department of Aeronautics and Astronautics, and senior author Nicholas Roy, an MIT professor of aeronautics and astronautics and a member of the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). The research will be presented at the International Conference on Robotics and Automation.

Generating graphs

To study motion planning, researchers often think about a robot’s environment like a graph, where a series of “edges,” or line segments, represent possible paths between a starting point and a goal.

Veys and her collaborators used a graph representation called the Canadian Traveler’s Problem (CTP), which draws its name from frustrated Canadian motorists who must turn back and find a new route when the road ahead is blocked by snow.

In a CTP, each edge of the graph has a weight associated with it, which represents how long that path will take to traverse, and a probability of how likely it is to be traversable. The goal in a CTP is to minimize travel time to the destination.

The researchers focused on how to automatically generate a CTP graph that effectively represents an uncertain environment.

“If we are navigating in an environment, it is possible that we have some information, so we are not just going in blind. While it isn’t a detailed navigation plan, it gives us a sense of what we are working with. The crux of this work is trying to capture that within the CTP graph,” adds Kurtz.

Their algorithm assumes this partial information — perhaps a satellite image — can be divided into specific areas (a lake might be one area, an open field another, etc.)

Each area has a probability that the robot can travel across it. For instance, it is more likely a nonaquatic robot can drive across a field than through a lake, so the probability for a field would be higher.

The algorithm uses this information to build an initial graph through open space, mapping out a conservative path that is slow but definitely traversable. Then it uses a metric the team developed to determine which edges, or shortcut paths through uncertain regions, should be added to the graph to cut down on the overall travel time.

Selecting shortcuts

By only selecting shortcuts that are likely to be traversable, the algorithm keeps the planning process from becoming needlessly complicated.

“The quality of the motion plan is dependent on the quality of graph. If that graph doesn’t have good paths in it, then the algorithm can’t give you a good plan,” Veys explains.

After testing the algorithm in more than 100 simulated experiments with increasingly complex environments, the researchers found that it could consistently outperform baseline methods that don’t consider probabilities. They also tested it using an aerial campus map of MIT to show that it could be effective in real-world, urban environments.

In the future, they want to enhance the algorithm so it can work in more than two dimensions, which could enable its use for complicated robotic manipulation problems. They are also interested in studying the mismatch between CTP graphs and the real-world environments those graphs represent.

“Robots that operate in the real world are plagued by uncertainty, whether in the available sensor data, prior knowledge about the environment, or about how other agents will behave. Unfortunately, dealing with these uncertainties incurs a high computational cost,” says Seth Hutchinson, professor and KUKA Chair for Robotics in the School of Interactive Computing at Georgia Tech, who was not involved with this research. “This work addresses these issues by proposing a clever approximation scheme that can be used to efficiently compute uncertainty-tolerant plans.”

This research was funded, in part, by the U.S. Army Research Labs under the Distributed Collaborative Intelligent Systems and Technologies Collaborative Research Alliance and by the Joseph T. Corso and Lily Corso Graduate Fellowship.

Study finds workers misjudge wage markets

Thu, 03/14/2024 - 12:00am

Many employees believe their counterparts at other firms make less in salary than is actually the case — an assumption that costs them money, according to a study co-authored by MIT scholars.

“Workers wrongly anchor their beliefs about outside options on their current wage,” says MIT economist Simon Jäger, co-author of a newly published paper detailing the study’s results.

As a top-line figure, the study indicates that workers who would experience a 10 percent wage increase by switching firms only expect a 1 percent wage increase instead, leading them to earn less than they otherwise might.

That is one of multiple related findings in the study, which also shows that workers in lower-paying firms are highly susceptible to underestimating wages at other companies; and that giving workers correct information about the salary structure in their industry makes them more likely to declare that they intend to leave their current jobs.

The study also has implications for further economics research, since economists’ job-search models generally assume workers have accurate salary information about their industries. The study was performed using data from Germany, although it quite likely applies to other countries as well.

“Misperceptions about outside options have substantial consequences on wages,” says Nina Roussille, an economist at MIT and also a co-author of the paper. “The intuition is simple: If low-wage workers do not know that they could make more elsewhere, then these workers stay put in low-wage firms. In turn, these low-wage firms do not feel the competitive pressure from the external labor market to raise their wages.”

The paper, “Worker Beliefs about Outside Options,” appears in advance online form in the Quarterly Journal of Economics. The authors are Jäger, the Silverman Family Career Development Associate Professor in MIT’s Department of Economics; Christopher Roth, a professor of economics at the University of Cologne; Roussille, an assistant professor in MIT’s Department of Economics; and Benjamin Schoefer, an associate professor of economics at the University of California at Berkeley.

Updating beliefs

To conduct the study, the researchers incorporated a survey module into the Innovation Sample of the German Socio-Economic Panel, an annual survey of a representative sample of the German population. They used their survey questions to find out the nature of worker beliefs about outside employment opportunities. The scholars then linked these findings to actual job and salary data collected from the German government’s Institute for Employment Research (IAB), with the prior consent of 558 survey respondents.

Linking those two data sources allowed the scholars to quantify the mismatch between what workers believe about industry-wide salaries, and what wages are in reality. One good piece of evidence on the compression of those beliefs is that about 56 percent of respondents believe they have a salary in between the 40th and 60th percentiles among comparable workers.

The scholars then added another element to the research project. They conducted an online experiment with 2,448 participants, giving these workers correct information about salaries at other companies, and then measuring the employees’ intention to find other job opportunities, among other things.

By adding this layer to the study, the scholars found that a 10 percentage point increase in the belief about salaries at other firms leads to a 2.6 percentage point increase in a worker intending to leave their present firm.

“This updating of beliefs causes workers to adjust their job search and wage negotiation intentions,” Roussille observes.

While the exact circumstances in every job market may vary somewhat, the researchers think the basic research findings from Germany could well apply in many other places.

“We are confident the results are representative of the German labor market,” Jäger says. “Of course, the German labor market may differ from, say, the U.S. labor market. Our intuition, though, is that, if anything, misperceptions would be even more consequential in a country like the U.S. where wages are more unequal than in Europe.”

Moreover, he adds, the recent dynamics of the U.S. job market during the Covid-19 pandemic, when many workers searched for new work and ended up in higher-paying jobs, is “consistent with the idea that workers had been stuck in low-paying jobs for a long time without realizing that there may have been better opportunities elsewhere.”

Data informing theory

The findings of Jäger, Roth, Roussille, Schoefer stand in contrast to established economic theory in this area, which has often worked from the expectation that employees have an accurate perception of industry wages and make decisions on that basis.

Roussille says the feedback the scholars have received from economics colleagues has been favorable, since other economists perceive “an opportunity to better tailor our models to reality,” as she puts it. “This follows a broader trend in economics in the past 20 to 30 years: The combination of better data collection and access with greater computing power has allowed the field to challenge longstanding but untested assumptions, learn from new empirical evidence, and build more realistic models.”

The findings have also encouraged the scholars to explore the topic further, especially by examining what the state of industry-wide wage knowledge is among employers.

“One natural follow-up to this project would be to better understand the firm side,” Jäger says. “Are firms aware of these misperceptions? Do they also hold inaccurate beliefs about the wages at their competitors?”

To this end, the researchers have already conducted a survey of managers on this topic, and intend to pursue further related work.

Support for the research was provided, in part, by the Sloan Foundation’s Working Longer Program; the Stiftung Grundeinkommen (Basic Income Foundation); and the Deutsche Forschungsgemeinschaft (German Research Foundation) under Germany’s Excellence Strategy.

“Imagine it, build it” at MIT

Thu, 03/14/2024 - 12:00am

MIT class 2.679 (Electronics for Mechanical Systems II) offers a sort of alchemy that transforms students from consumers of knowledge to explorers and innovators, and equips them with a range of important new tools at their disposal, students say.

“Topics which could otherwise feel intimidating are well-scoped each week so that students come out knowing not only what a concept is, but why it’s useful and how to actually implement it,” says graduating senior Audrey Chen. “I could consistently come in with no background and come out with practical experience I could use in future projects. I’d describe the class as a series of small crash courses [each of which] answers, simply, ‘what do I need to know to do or use this thing?’”

The course takes students through the process of design, fabrication, and assembly of a printed circuit board (PCB). Ultimately, that process, which has twists and turns depending on each student’s project idea, culminates in incorporating the PCB into a device — in a sense animating that device to perform a certain function.

“The design intent of 2.679 is to empower students to ‘imagine it, build it,’” says Tonio Buonassisi, professor of mechanical engineering. "Between those two is a universe, and the purpose of this class is to aid aspiring engineers to bridge that gap.”

Senior Jessica Lam marvels at how much she learned in the course over its one short semester, attributing that flood of education to the class labs being “incredibly well-structured.”

“I’ve found that in a lot of other labs and project-based classes, they throw a lot of information at you at once with the expectation that you already have some experience with certain software or hardware, and most of it is scaffolded and feels like a black box,” without much understanding of what is actually happening, Lam says. “In 2.679, Steve Banzaert has a better understanding of what we already know and how to build on that.”

After taking 2.679, she says she feels “a lot more confident in designing electrical systems, and I have a more comprehensive understanding of how to integrate mechanical systems and electronics.”

Banzaert, technical instructor for the course, says the class is designed to guide students along their own chosen paths of discovery, showing them that they are able to address the challenges they encounter along the way.

“Every semester we get to see really lovely examples of growth, not just in the course material but, in the best cases, in students’ understanding of what they’re really capable of,” he says.

Chen, a mechanical engineering major who is graduating early to start a position as a hardware project manager at Formlabs, agrees that the class did just that.

“Students are given tremendous freedom to pick their own final projects, allowing them to explore topics which are of special interest to them. And because each project is unique, there is less pressure to ‘perform’ in a traditional sense,” she says. “Rather, each student is learning different skills and is encouraged to get as far along with the project they choose as possible. Steve emphasized that the scope of our projects would inevitably change, because at the start you simply don’t yet know what you don’t know, and that’s totally okay!”

Banzaert says, “We try to make it very clear that, yes, we are talking about important general concepts in theory and analysis, but that’s because they are tools that engineers use to solve problems. I think maybe this focus helps remind the students of what got them here in the first place — that the reason you’re an engineer is because there’s something about the world you wish was better, that you’re the person to do it (or at least help), and, if you want to do it well, you’re going to have to learn a bunch of things so you have more tools in your toolbox.”

Senior Yasin Hamed designed a car in the class that uses computer vision to follow along a black line. The car has an attached camera that captures images and relays them to a Raspberry Pi computer that is also attached to the car. Processing the images in real time allows the car to locate the black line and turn or go straight while controlling the car’s speed.

Although Hamed, who is majoring in mechanical engineering with a minor in computer science, had built another similar system in a previous class, he says the focus in the prior class was on the software. With his 2.679 car project, he learned about “the underlying foundation,” meaning “the design of the power electronics and control circuitry which is necessary for everything else to work.”

“I derived much of the ‘enlightenment’ from this class from the little electronic bits and pieces of information I picked up along the course of the class, like learning/practicing soldering, understand how to use integrated circuits, learning how to design a PCB, etc.,” he says. “It was the collection of all of these things that benefited me the most.”

Jordan Parker-Ashe, also a senior, appreciated how 2.679 combined lessons about electronics with research and presentations from Buonassisi’s lab. “It’s great seeing engineering applied in research,” she says.

Although many of the skills she learned in the course were new to her, one was “an old foe,” she says, that 2.679 allowed her to befriend. Parker-Ashe, who is majoring in nuclear engineering, had used a computer vision program called OpenCV in her first Undergraduate Research Opportunities Program project as a first-year undergraduate.

“It was the hardest thing ever, and it really felt like an insurmountable obstacle then,” she says. “Now, to be using OpenCV in labs and homework effortlessly — It was a very full-circle moment.”

She says the class has opened up a whole new field to her, with Banzaert having “directly inspired” her to also take class 6.131 (Power Electronics), “which has been life-changing,” she says.

“2.679 helped me believe in myself, which inspired me to take 6.131, a notorious electrical engineering capstone, which has made me realize that my future lies as a nuclear-electrical engineering engineer, not just a nuclear engineer,” Parker-Ashe says. “I want to pursue electrical engineering in my future, and that just wasn’t on the table beforehand.

“Not to mention that it’s opened the doors to very rich landscapes for project ideas, creating explorations, art, stepping into new roles in group projects, etc,” she says. "I'm so glad that I've been able to find opportunities in Course 2  that helped give me hands-on, applied engineering experience."

Pages