MIT Latest News
Dylan Cable’s goal is to uncover the inner workings of vital biological processes. This aim may be a bit closer now, as he has just earned a prestigious fellowship from the nonprofit Fannie and John Hertz Foundation.
Cable is one of 11 scholars from nine U.S. research universities chosen from a pool of 840 applicants. The award, announced this week, provides recipients with up to $250,000 for up to five years. In addition to funding, the researchers are free from many of the constraints that other fellowships entail.
In making the announcement, David Galas, chairman of the Hertz Foundation’s board of directors, noted, “It is increasingly challenging to get funding for truly scientific research, but it is even more so for young researchers to pursue their own ideas.” The foundation supports “groundbreaking applied science with real-world benefits for all humanity.”
A first-year PhD student in MIT’s Department of Electrical Engineering and Computer Science, Cable grew up in Chicago, where he developed a deep passion for mathematics. This passion became his undergraduate major at Stanford University, where his exploration of neuroscience and biology convinced him that life science problems are the most important issues to solve.
Cable’s current work involves improving physical methods for biological data collection and creating mathematical methods for biological data analysis. He regards data collection and analysis as inseparable and believes they must be interwoven to achieve deep knowledge of biological problems.
To win the Hertz Fellowship, Cable’s application proceeded through a rigorous selection process in which a candidate must demonstrate a “background of leadership in research, achievement, and proven technical understanding of their field” in addition to academic excellence. During the unique interview process, the candidate must also demonstrate a capacity to think creatively and to “push the traditional boundaries of research.”
Cable co-authored a recent article in the Journal of Neuroscience that examines whether single neuron recordings can be recorded from humans using functional magnetic resonance imaging (fMRI), a question with major implications for studying the human brain.
As he pursues his academic path, Cable joins approximately 20 other MIT graduate students completing their tenure as Hertz Fellows. He also has access to unique learning experiences with alumni scholars, entrepreneurs, and scientific leaders in the broader Hertz community. Former fellows have been honored with the Nobel Prize, the National Medal of Science, the Turing Award, the Breakthrough Prize, and the MacArthur Fellowship (“Genius Grant”).
“The future of work will be determined by who yields power and for what purposes. We are in a moment of great transition — we have an opportunity to imagine what a new social contract can be,” said Sarita Gupta, executive director of Jobs with Justice and co-director of Caring Across Generations as she kicked off last Friday’s launch of J-PAL North America’s Work of the Future Initiative.
Gupta opened the confercence with a powerful call to action for participants to shift the narrative around “the future of work.” The newest initiative from J-PAL North America, a research center in MIT’s Department of Economics, the Work of the Future Initiative seeks to identify effective, evidence-based strategies that increase opportunity, reduce disparities, and help all workers navigate the work of the future.
Millions of workers throughout the industrialized and developing worlds could be affected by automation, rising inequality, stagnating educational attainment, and other labor market trends in the coming decades. Many workers lack access to jobs that pay living wages, have jobs with insufficient benefits or protections, or lack the necessary skills or education to progress within their industries in the face of technological change.
By spurring research on effective ways to help workers thrive in today’s changing labor market, the Work of the Future Initiative aims to center worker voices and create a more equitable future of work. The conference addressed a number of big questions, including: How can the future of work be made more equitable, efficient, and just?
“J-PAL North America’s Work of the Future Initiative was launched to catalyze rigorous research on these urgent questions,” explained David Autor, the Ford Professor of Economics at MIT and co-chair of the new initiative.
Autor also serves as vice-chair of the Institute’s complementary Work of the Future Task Force, a recently-launched group of MIT faculty and researchers exploring how emerging technologies are changing the nature of human work and what types of education and skills will enable humans to thrive in the digital economy.
The Initiative’s academic leadership, including Autor, co-chairs Matthew Notowidigdo of Northwestern University, and J-PAL Scientific Director Lawrence Katz of Harvard University, recognized that across the country, policymakers, industry leaders, and social service providers are actively seeking solutions to labor market challenges.
Many well-intentioned, potentially effective ideas remain untested, however, leaving policymakers without the necessary evidence to assess what will be helpful, neutral, or harmful. Too often, academic researchers, government agencies, and nonprofit and industry leaders are working on these critical problems in isolation, and don’t have the time or resources to tap into each other’s expertise.
J-PAL’s newest initiative seeks to fill this gap by generating new research to help answer these important questions. It will catalyze this kind of rigorous, actionable evidence through an innovation competition model and a researcher-facing request for proposals (RFP).
The innovation competition is currently accepting promising research proposals from practitioners across the country, and will work with selected partners to develop a feasible, rigorous evaluation of a program or policy focused on the future of work.
Selected applicants will receive technical support from J-PAL staff, flexible funding to get an evaluation off the ground, and access to J-PAL’s network of leading academic researchers to help them design and implement randomized evaluations of their programs.
Evelyn Diaz, president of Heartland Alliance and a panelist at the kick-off event, explained why this kind of rigorous evaluation is critical to an organization’s success. “There is a fear of failure about evaluation, and we need to change the narrative,” Diaz said. “The focus should instead be on how we are learning.”
Those seeking to learn more about the competition are encouraged to sign up for J-PAL's informational webinar on June 26. Through the competition, along with a bi-annual, researcher-facing RFP, the initiative aims to generate actionable research on questions related to the future of work.
Meawhile, with conferences like last Friday’s kick-off event, the initiative will also serve as a convener to bring together leading voices in the future of work space. At the kick-off, participants from academic institutions, nonprofits, philanthropies, and the private sector gathered to share insights, learn from each other’s different experiences, and brainstorm solutions to complex research questions.
Event highlights included a number of engaging, interdisciplinary panels on challenges and opportunities related to the work of the future.
Gupta and Katz, for example, participated in a lively discussion with Abigail Wozniak from the Federal Reserve Bank of Minneapolis on how to shift narratives around the future of work.
Later in the day, Notowidigdo presented key findings from his recent research agenda on the Work of the Future, co-authored by Autor and Northwestern University graduate student Anran Li, and an interdisciplinary panel of industry and nonprofit leaders and academic researchers provided thoughtful commentary on the research agenda.
Jed Kolko, chief economist at Indeed, echoed the review paper’s call for more rigorous research on these topics. “There is a lot of uncertainty about the effect of automation technology on employment. Setting up experiments that will be able to measure those effects is critical.”
David Autor also presented innovative research on how work has — and hasn’t — changed over time, and the implications of this research for worker well-being, and an interdisciplinary panel of researchers and practitioners discussed how they formed mutually beneficial research-practitioner partnerships.
To wrap up the day, J-PAL North America Executive Director Mary Ann Bates moderated a wide-ranging panel on the changing nature of work in the United States that included Katz, J-PAL affiliate Damon Jones, and Julie Gehrki, vice president of philanthropy at the Walmart Foundation.
Bates’ opening remarks on the motivating principle behind the initiative set the tone for the rest of the day’s discussions. “The reason why we care about these topics is because of people,” she said.
In February, the Institute established five working groups to generate ideas for different components of the structure and operation of the new MIT Stephen A. Schwarzman College of Computing. Three community forums for all five working groups were held Wednesday, April 17, and Thursday, April 18. Troy Van Voorhis, the Haslam and Dewey Professor of Chemistry, and Srini Devadas, the Webster Professor of Electrical Engineering and Computer Science, are co-chairs of the Working Group on Curricula and Degrees, which is charged with studying how to develop new curricula for the college, what degrees should be shifted from their existing departments into the college, what new degrees or other credentials might be created, and how to design dual-degree programs with existing departments. MIT News checked in with Van Voorhis and Devadas to find out about the group’s goals, processes, and progress so far.
Q: What process has your working group undergone in preparing your report, and how many people have been involved?
A: We have about 15 members in the committee not including the co-chairs. All schools are represented, as are staff and students. We meet each week for an hour and there are email discussions between meetings. We started with naming ourselves CoC2 (College of Computing Committee on Curricula) and drafting an educational mission for the college. We have been refining our mission statement throughout our discussions.
We have had extensive discussions on how the college can provide a broad funnel for undergraduate and graduate students interested in computing by offering various types of credentials, including minors, joint degree programs, and certificates. During these meetings, we have discussed the pros and cons of current credentials, and members have proposed new variants that might better serve the needs of students.
At the graduate level, we discussed the Business Analytics Certificate in Sloan as an example of a structure that we might want to replicate in the college, but with a focus on computation. We have begun writing our report that is due at the end of the semester.
Q: Could you describe any areas that participants in the process have readily agreed on, or others that have turned out to be contentious?
A: We have focused primarily on pedagogical aspects thus far, and not on operational aspects, for example: faculty receiving credit for teaching courses jointly across the college and other schools; whether the college is responsible for all joint majors, as opposed to departments, etc. This is largely because these questions are clearly dependent on the eventual structure of the college and the responsibilities of faculty who are primary in the college and those that are affiliated. Given our timeline, it makes sense to visit these questions in a holistic manner after all the committees have written their reports.
Our discussions have been always informative and often passionate, but not contentious in the least. As an example, we were able to draft an educational mission that everyone largely agreed with in short order, and the ongoing refinement has been about getting nuances “right.” Our meeting with the Societal Impact committee was productive and impacted our mission statement, and will impact our report.
Q: What do you see as the next steps once you finish your working group's report?
A: Provost [Martin] Schmidt, Dean [Dan] Huttenlocher, and the administration will determine next steps. First, the organizational structure of the college needs to be determined — in other words, what departments are in the college, and more important from our committee’s standpoint, what degree programs will be the college’s responsibility. A very important set of decisions relates to faculty appointments in the college and how credit for teaching classes in the college is assigned by departments within and outside the college.
There is some work that can proceed in parallel — for example, a new committee could engage the Committee on Curricula to discuss potential flexibility in the current restriction of “at most two courses in a minor can be used to satisfy a major requirement.” The exploration of degree programs that offer a truly integrated experience across computation and another discipline is another possibility.
When Katie O’Nell’s high school biology teacher showed a NOVA video on epigenetics after the AP exam, he was mostly trying to fill time. But for O’Nell, the video sparked a whole new area of curiosity.
She was fascinated by the idea that certain genes could be turned on and off, controlling what traits or processes were expressed without actually editing the genetic code itself. She was further excited about what this process could mean for the human mind.
But upon starting at MIT, she realized that she was less interested in the cellular level of neuroscience and more fascinated by bigger questions, such as, what makes certain people generous toward certain others? What’s the neuroscience behind morality?
“College is a time you can learn about anything you want, and what I want to know is why humans are really, really wacky,” she says. “We’re dumb, we make super irrational decisions, it makes no sense. Sometimes it’s beautiful, sometimes it’s awful.”
O’Nell, a senior majoring in brain and cognitive sciences, is one of five MIT students to have received a Marshall Scholarship this year. Her quest to understand the intricacies of the wacky human brain will not be limited to any one continent. She will be using the funding to earn her master’s in experimental psychology at Oxford University.
Chocolate milk and the mouse brain
O’Nell’s first neuroscience-related research experience at MIT took place during her sophomore and junior year, in the lab of Institute Professor Ann Graybiel.
The research studied the neurological components of risk-vs-reward decision making, using a key ingredient: chocolate milk. In the experiments, mice were given two options — they could go toward the richer, sweeter chocolate milk, but they would also have to endure a brighter light. Or, they could go toward a more watered-down chocolate milk, with the benefit of a softer light. All the while, a fluorescence microscope tracked when certain cell types were being activated.
“I think that’s probably the closest thing I’ve ever had to a spiritual experience … watching this mouse in this maze deciding what to do, and watching the cells light up on the screen. You can see single-cell evidence of cognition going on. That’s just the coolest thing,” she says.
In her junior spring, O’Nell delved even deeper into questions of morality in the lab of Professor Rebecca Saxe. Her research there centers on how the human brain parses people’s identities and emotional states from their faces alone, and how those computations are related to each other. Part of what interests O’Nell is the fact that we are constantly making decisions, about ourselves and others, with limited information.
“We’re always solving under uncertainty,” she says. “And our brain does it so well, in so many ways.”
Outside of class, O’Nell has no shortage of things to do. For starters, she has been serving as an associate advisor for a first-year seminar since the fall of her sophomore year.
“Basically it’s my job to sit in on a seminar and bully them into not taking seven classes at a time, and reminding them that yes, your first 8.01 exam is tomorrow,” she says with a laugh.
She has also continued an activity she was passionate about in high school — Model United Nations. One of the most fun parts for her is serving on the Historical Crisis Committee, in which delegates must try to figure out a way to solve a real historical problem, like the Cuban Missile Crisis or the French and Indian War.
“This year they failed and the world was a nuclear wasteland,” she says. “Last year, I don’t entirely know how this happened, but France decided that they wanted to abandon the North American theater entirely and just took over all of Britain’s holdings in India.”
She’s also part of an MIT program called the Addir Interfaith Fellowship, in which a small group of people meet each week and discuss a topic related to religion and spirituality. Before joining, she didn’t think it was something she’d be interested in — but after being placed in a first-year class about science and spirituality, she has found discussing religion to be really stimulating. She’s been a part of the group ever since.
O’Nell has also been heavily involved in writing and producing a Mystery Dinner Theater for Campus Preview Weekend, on behalf of her living group J Entry, in MacGregor House. The plot, generally, is MIT-themed — a physics professor might get killed by a swarm of CRISPR nanobots, for instance. When she’s not cooking up murder mysteries, she might be running SAT classes for high school students, playing piano, reading, or spending time with friends. Or, when she needs to go grocery shopping, she’ll be stopping by the Trader Joe’s on Boylston Avenue, as an excuse to visit the Boston Public Library across the street.
Quite excited for the future
O’Nell is excited that the Marshall Scholarship will enable her to live in the country that produced so many of the books she cherished as a kid, like “The Hobbit.” She’s also thrilled to further her research there. However, she jokes that she still needs to get some of the lingo down.
“I need to learn how to use the word ‘quite’ correctly. Because I overuse it in the American way,” she says.
Her master’s research will largely expand on the principles she’s been examining in the Saxe lab. Questions of morality, processing, and social interaction are where she aims to focus her attention.
“My master’s project is going to be basically taking a look at whether how difficult it is for you to determine someone else’s facial expression changes how generous you are with people,” she explains.
After that, she hopes to follow the standard research track of earning a PhD, doing postdoctoral research, and then entering academia as a professor and researcher. Teaching and researching, she says, are two of her favorite things — she’s excited to have the chance to do both at the same time. But that’s a few years ahead. Right now, she hopes to use her time in England to learn all she can about the deeper functions of the brain, with or without chocolate milk.
Beginning in 2020, the timing of MIT’s Hooding and Commencement ceremonies will move from June to late May in most years. The 2020 ceremonies will be held on Thursday, May 28 and Friday, May 29. (Projected dates for doctoral hooding and Commencement ceremonies that will occur between 2021 and 2025 are available via the MIT Registrar.)
Faculty approved these updates to the academic calendar at the faculty meeting on April 17. The vote marked the culmination of months of community engagement and input-gathering conducted by the Office of the Vice Chancellor.
To accommodate the change, the Patriots’ Day student holiday will be reduced from four days to a three-day-weekend. The timing of the Independent Activities Period (IAP) will remain the same, and Registration Day will occur on the last day of IAP (Friday). The first day of spring classes will be on a Monday, and the last day of classes will be on a Tuesday. There will be four days of finals that straddle the weekend, but there will be an additional reading day. The resulting format will include two reading days, one exam day (Friday), two reading days, three exam days (Monday-Wednesday). The number of spring term teaching days remains at 65.
“The goal was to make adjustments to our academic calendar to enable a longer summer period and do so in a way that preserves what we heard was most important to the community,” said Vice Chancellor Ian Waitz, who, along with Registrar and Senior Associate Dean Mary Callahan, worked with hundreds of community members to devise a solution that best accommodates students, faculty, and staff. “With this change, we are helping the students who have to delay employment and internship opportunities, or extend their housing rental agreements, because of a June graduation date.”
Waitz, who is also the Jerome C. Hunsaker Professor of Aeronautics and Astronautics, noted that the new timing will assist faculty who currently need to put off early summer research and conference travel. In addition, the change will allow MIT to start maintenance and capital projects in residence halls and classrooms, and launch campus summer programs, sooner than is possible now.
Given the complex nature of the academic calendar, Waitz and Callahan acknowledged that the three proposals they considered, and the final version that was approved, come with positives and negatives. Even though the majority of people who provided feedback preferred the version that was approved, there are some aspects of the new format that can be further improved through proactive problem-solving and clear communication. For example, a group of staff is meeting now to develop different accommodation strategies for graduating students and their families, who may face conflicts with religious holidays or athletic schedules as a result of the new ceremony date.
“We will help the community understand and work through the implementation of this change recognizing that an academic calendar shift may create some challenges,” said Callahan.
Throughout the fall and spring, Callahan and Waitz engaged faculty by meeting with dean's group, academic council, undergraduate officers, heads of house, the OME Faculty Advisory Committee, the Committee on Graduate Programs, the Faculty Policy Committee, the Committee on the Undergraduate Program, and Committee on Academic Performance. They worked with students from the Undergraduate Association and the Graduate Student Council as well as within the faculty governance structure, the Undergraduate and Graduate academic administrators, staff from the Office of Minority Education (OME), Global Education, and Student Support Services (S3), and consulted the faculty, students, and staff members of the Commencement Committee. In addition, the community had an opportunity to provide feedback via an online survey and a student forum hosted by the Office of the Vice Chancellor.
For additional information, please refer to these Frequently Asked Questions.
Lockheed Martin and MIT International Science and Technology Initiatives (MISTI) have announced the creation of the MIT-Lockheed Martin Seed Fund to promote early-stage collaborations between MIT faculty and researchers with universities and public research institutions in Israel. The seed fund will also take place in Germany, and additional countries will be considered after the pilot year of 2019.
The MIT-Lockheed Martin Seed Fund, to be sponsored by Lockheed Martin with more than $150,000, includes two to four projects for Israel and two to four projects for Germany. MIT will administrate the fund within the MIT International Science and Technology Initiatives program in the Center for International Studies. This funding may be used for travel, meeting, and workshop costs, inclusive of visits to Lockheed Martin and MIT facilities in the U.S. Furthermore, the seed fund includes one MIT program student internship in Israel as part of the MIT-Israel program and one MIT student internship in Germany as part of the MIT-Germany program.
For the inaugural year, the seed fund will focus on proposals that fit within Lockheed Martin’s Advanced Manufacturing priorities to identify emerging innovative technologies around but not limited to:
• manufacturing process control;
• modeling of materials and processes;
• novel materials for extreme environments; and
• automation of the "Factory of the Future."
This collaboration brings the ability to align projects around Lockheed Martin’s areas of technology interest and interface with top global universities under a structured program which may lead to sponsored research under a separate agreement. Collaborating faculty will have the opportunity move forward their joint projects as well as engage with Lockheed Martin facilities in the U.S. and Israel.
The new Lockheed Martin-MISTI initiative joins the collaboration made in recent years with Israel’s Ministry of Education, Ministry of Science and Technology, and the Rashi Foundation to promote STEM-related programs from kindergarten throughout high school to higher education.
Headquartered in Bethesda, Maryland, Lockheed Martin is a global security and aerospace company that employs approximately 105,000 people worldwide and is principally engaged in the research, design, development, manufacture, integration, and sustainment of advanced technology systems, products, and services.
MIT International Science and Technology Initiatives creates applied international learning opportunities for MIT students that increase their ability to understand and address real-world problems and bolsters MIT’s research mission by promoting collaborations between MIT faculty members and their counterparts abroad. Since 2008, MISTI’s Global Seed Fund program has awarded $17.7 million to over 800 faculty projects. MISTI is housed within the MIT School of Humanities, Arts, and Social Sciences.
The work of a science writer, including this one, includes reading journal papers filled with specialized technical terminology, and figuring out how to explain their contents in language that readers without a scientific background can understand.
Now, a team of scientists at MIT and elsewhere has developed a neural network, a form of artificial intelligence (AI), that can do much the same thing, at least to a limited extent: It can read scientific papers and render a plain-English summary in a sentence or two.
Even in this limited form, such a neural network could be useful for helping editors, writers, and scientists scan a large number of papers to get a preliminary sense of what they’re about. But the approach the team developed could also find applications in a variety of other areas besides language processing, including machine translation and speech recognition.
The work is described in the journal Transactions of the Association for Computational Linguistics, in a paper by Rumen Dangovski and Li Jing, both MIT graduate students; Marin Soljačić, a professor of physics at MIT; Preslav Nakov, a senior scientist at the Qatar Computing Research Institute, HBKU; and Mićo Tatalović, a former Knight Science Journalism fellow at MIT and a former editor at New Scientist magazine.
From AI for physics to natural language
The work came about as a result of an unrelated project, which involved developing new artificial intelligence approaches based on neural networks, aimed at tackling certain thorny problems in physics. However, the researchers soon realized that the same approach could be used to address other difficult computational problems, including natural language processing, in ways that might outperform existing neural network systems.
“We have been doing various kinds of work in AI for a few years now,” Soljačić says. “We use AI to help with our research, basically to do physics better. And as we got to be more familiar with AI, we would notice that every once in a while there is an opportunity to add to the field of AI because of something that we know from physics — a certain mathematical construct or a certain law in physics. We noticed that hey, if we use that, it could actually help with this or that particular AI algorithm.”
This approach could be useful in a variety of specific kinds of tasks, he says, but not all. “We can’t say this is useful for all of AI, but there are instances where we can use an insight from physics to improve on a given AI algorithm.”
Neural networks in general are an attempt to mimic the way humans learn certain new things: The computer examines many different examples and “learns” what the key underlying patterns are. Such systems are widely used for pattern recognition, such as learning to identify objects depicted in photos.
But neural networks in general have difficulty correlating information from a long string of data, such as is required in interpreting a research paper. Various tricks have been used to improve this capability, including techniques known as long short-term memory (LSTM) and gated recurrent units (GRU), but these still fall well short of what’s needed for real natural-language processing, the researchers say.
The team came up with an alternative system, which instead of being based on the multiplication of matrices, as most conventional neural networks are, is based on vectors rotating in a multidimensional space. The key concept is something they call a rotational unit of memory (RUM).
Essentially, the system represents each word in the text by a vector in multidimensional space — a line of a certain length pointing in a particular direction. Each subsequent word swings this vector in some direction, represented in a theoretical space that can ultimately have thousands of dimensions. At the end of the process, the final vector or set of vectors is translated back into its corresponding string of words.
“RUM helps neural networks to do two things very well,” Nakov says. “It helps them to remember better, and it enables them to recall information more accurately.”
After developing the RUM system to help with certain tough physics problems such as the behavior of light in complex engineered materials, “we realized one of the places where we thought this approach could be useful would be natural language processing,” says Soljačić, recalling a conversation with Tatalović, who noted that such a tool would be useful for his work as an editor trying to decide which papers to write about. Tatalović was at the time exploring AI in science journalism as his Knight fellowship project.
“And so we tried a few natural language processing tasks on it,” Soljačić says. “One that we tried was summarizing articles, and that seems to be working quite well.”
The proof is in the reading
As an example, they fed the same research paper through a conventional LSTM-based neural network and through their RUM-based system. The resulting summaries were dramatically different.
The LSTM system yielded this highly repetitive and fairly technical summary: “Baylisascariasis,” kills mice, has endangered the allegheny woodrat and has caused disease like blindness or severe consequences. This infection, termed “baylisascariasis,” kills mice, has endangered the allegheny woodrat and has caused disease like blindness or severe consequences. This infection, termed “baylisascariasis,” kills mice, has endangered the allegheny woodrat.
Based on the same paper, the RUM system produced a much more readable summary, and one that did not include the needless repetition of phrases: Urban raccoons may infect people more than previously assumed. 7 percent of surveyed individuals tested positive for raccoon roundworm antibodies. Over 90 percent of raccoons in Santa Barbara play host to this parasite.
Already, the RUM-based system has been expanded so it can “read” through entire research papers, not just the abstracts, to produce a summary of their contents. The researchers have even tried using the system on their own research paper describing these findings — the paper that this news story is attempting to summarize.
Here is the new neural network’s summary: Researchers have developed a new representation process on the rotational unit of RUM, a recurrent memory that can be used to solve a broad spectrum of the neural revolution in natural language processing.
It may not be elegant prose, but it does at least hit the key points of information.
Çağlar Gülçehre, a research scientist at the British AI company Deepmind Technologies, who was not involved in this work, says this research tackles an important problem in neural networks, having to do with relating pieces of information that are widely separated in time or space. “This problem has been a very fundamental issue in AI due to the necessity to do reasoning over long time-delays in sequence-prediction tasks,” he says. “Although I do not think this paper completely solves this problem, it shows promising results on the long-term dependency tasks such as question-answering, text summarization, and associative recall.”
Gülçehre adds, “Since the experiments conducted and model proposed in this paper are released as open-source on Github, as a result many researchers will be interested in trying it on their own tasks. … To be more specific, potentially the approach proposed in this paper can have very high impact on the fields of natural language processing and reinforcement learning, where the long-term dependencies are very crucial.”
The research received support from the Army Research Office, the National Science Foundation, the MIT-SenseTime Alliance on Artificial Intelligence, and the Semiconductor Research Corporation. The team also had help from the Science Daily website, whose articles were used in training some of the AI models in this research.
Vivienne Sze, an associate professor in the Department of Electrical Engineering and Computer Science (EECS), has received the 2018-2019 Harold E. Edgerton Faculty Achievement Award.
The award, announced at today’s MIT faculty meeting, commends Sze for “her seminal and highly regarded contributions in the critical areas of deep learning and low-power video coding, and for her educational successes and passion in championing women and under-represented minorities in her field."
Sze’s research involves the co-design of energy-aware signal processing algorithms and low-power circuit, architecture, and systems for a broad set of applications, including machine learning, computer vision, robotics, image processing, and video coding. She is currently working on projects focusing on autonomous navigation and embedded artificial intelligence (AI) for health-monitoring applications.
“In the domain of deep learning, [Sze] created the Eyeriss chip for accelerating deep learning algorithms, building a flexible architecture to handle different convolutional shapes,” the Edgerton Faculty Award selection committee said in announcing its decision. “Eyeriss is also the first deep neural network accelerator to exploit data statistics of the network to further reduce energy consumption twofold, a substantial accomplishment in this field.” In addition, the committee noted that Sze’s work on High Efficiency Video Coding (HEVC) influenced the development of standards now in widespread use, including the improvements that provide the enabling technology for video-accessing on iPhones and Android phones worldwide.
Sze’s educational contributions include a popular conference tutorial on hardware architectures for deep neural networks, which she has turned into a regularly offered MIT subject that can be used to satisfy the EECS doctoral qualifying procedures. In addition, students praise “her ability to connect theory and practice through enjoyable and helpful lectures,” the selection committee noted.
Finally, the committee acknowledged Sze’s efforts in promoting the inclusion and advancement of women and under-represented minorities in the field. Most recently, she served as a technical co-chair of Rising Stars in EECS 2018, which brought 76 of the world’s top women graduate students and postdoctoral researchers to MIT for an intensive two-day workshop on academic careers. (As a PhD student, Sze participated in the inaugural Rising Stars in EECS workshop, held at MIT in 2012.)
Sze received a bachelor’s degree with honors from the University of Toronto and a master’s degree and PhD from MIT, all in electrical engineering. From 2010 to 2013, she was a member of the technical staff in the R&D Center at Texas Instruments (TI), where she designed low-power algorithms and architectures for video coding. In that role, she represented TI on the Joint Collaborative Team on Video Coding during the development of HEVC, which later received an Engineering Emmy Award. Sze returned to MIT in 2013 as an assistant professor of electrical engineering and computer science and was promoted to associate professor without tenure in July 2017. She is also a principal investigator in the Research Lab of Electronics, has co-authored more than 60 publications in proceedings of refereed conferences and journals, and holds more than 25 issued patents.
Other honors include a 2018 Google Faculty Research Award, a 2018 Facebook Faculty Award, a 2018 Qualcomm Faculty Award, and a 2016 Young Investigator Research Program Award from the Air Force Office of Scientific Research, among others. In 2011, she received MIT’s Jin-Au Kong Outstanding Doctoral Thesis Prize in Electrical Engineering.
The annual Edgerton Faculty Award was established in 1982 as a tribute to Institute Professor Emeritus Harold E. Edgerton in recognition of his active support of junior faculty members. Each year, a committee presents the award to one or more non-tenured faculty members to recognize exceptional contributions in research, teaching, and service.
The 2018-2019 Edgerton Award Selection Committee was chaired by Professor Ian Hunter, the George N. Hatsopoulos Professor in Thermodynamics in the Department of Mechanical Engineering. Committee members included Catherine Drennan, professor of chemistry and biology; Jay Scheib, Class of 1949 Professor of Theater; Antoinette Schoar, the Michael M. Koerner (1949) Professor of Finance and Entrepreneurship at the MIT Sloan School of Management; and Andrew Scott, professor of architecture and urbanism.
Before computers, no sane person would have set out to count gender pronouns in 4,000 novels, but the results can be revealing, as MIT’s new digital humanities program recently discovered.
Launched with a $1.3 million grant from the Andrew W. Mellon Foundation, the Program in Digital Humanities brings computation together with humanities research, with the goal of building a community “fluent in both languages,” says Michael Scott Cuthbert, associate professor of music, Music21 inventor, and director of digital humanities at MIT.
“In the past, it has been somewhat rare, and extremely rare beyond MIT, for humanists to be fully equipped to frame questions in ways that are easy to put in computer science terms, and equally rare for computer scientists to be deeply educated in humanities research. There has been a communications gap,” Cuthbert says. “That's the genesis of this new approach to computation in humanities.”
Educating bilinguals: students fluent in the humanities and computation
While traditional digital humanities programs attempt to provide humanities scholars with some computational skills, the situation at MIT is different: Most MIT students already have or are learning basic programming skills, and all MIT undergraduates also take some humanities classes. Cuthbert believes this difference will make MIT’s program a great success.
“What we have that's an amazing opportunity is a large number of people who love building things with computers and want to connect those to their interests and make an impact,” he says. “Our students very much want to change the world.”
They can do that — even as first-year students — because humanities research has many open questions that can be solved with just six months or a year of programming skills, he says.
“The wonderful thing we can do is implement a lot from scratch because we have the programming skills to do that,” says Stephan Risi, one of two postdocs who works in what the students informally call the “Digital Humanities Lab,” or “DH Lab” for short. This gives the MIT researchers more latitude to explore new questions as they arise. “We’re not bound by software others have produced.”
A novel research project
To illustrate the kind of work the lab can do, the program enlisted a team of 24 students (mostly first-years) through the MIT Undergraduate Research Opportunities Program (UROP) to study gender representation in 19th century English literature. The team assembled metadata, applied grammar-parsing tools, did web scraping, wrote analysis tools, and ultimately examined 4,217 books — a total of 326.9 million words.
One interesting finding from the "Gender/Novels" experiment was that — regardless of the sex of the author and "no matter how we cut the data,” as Cuthbert says — roughly two-thirds of all male pronouns were in the subject position, whereas women were more often the object of the sentence. What these new data tell us — about men, women, and society — is up to human scholars to decide, but this project provides a window into the ways computational work can support humanities research.
Detecting research with high social value
This first project also illustrates the pedagogical benefits of working in the lab.
“One of the interesting things about the lab is it's hard to sift through which ideas have merit,” says lab UROP and first-year student Dina Atia, contrasting the humanities research to her work in science, technology, engineering, and math (STEM) fields. “Most STEM research is very fact-based but can lack important social takeaways.”
Fellow UROP and first-year student Ifeoluwapo Ademolu-Odeneye says she enjoyed the opportunity to put her computer skills to work outside the classroom. “I have developed a lot as a computer scientist doing this,” she says, adding that she has also learned to apply critical thinking skills to make decisions about the humanities content. “At first, I asked Professor Cuthbert about everything. Later he threw questions back at us, which has been good for developing as a researcher myself.”
First-year student Mayowa Songonuga, who just started her UROP in the lab this spring and is working on a new project — The History of Computing at MIT — agreed that the hands-on work is very valuable. “There is more to it than just the technology,” she says. “I haven't had the chance to research something like this before.”
The productive swerve in research
While the UROP students were designing algorithms and building a website, they also read and analyzed 19th century English literature and tackled questions such as how to teach the computer the difference between a novel and a travel log. The lab intentionally fosters this dual-stream process, Cuthbert says, because it provides rich opportunities to change the direction of research to follow some newly discovered path.
This ability to make what Cuthbert calls “a productive swerve” is often critical to fruitful research, but has been hampered in the digital humanities to date because complex digital projects are too often done by computational experts at a remove from the humanities scholar.
Students collaborate with leading humanities scholars
To further entwine the disciplines, the program next plans to bring humanities faculty on board for joint projects with students. In 2019-20, associate professor of literature Sandy Alexandre and professor of political science Evan Lieberman will be devoting six hours a week to the lab, teaching students about their research while learning some computational methods themselves.
An added benefit of this collaboration is that it should make the programming work less demanding, Cuthbert says, because creating a simple user interface can be extremely time-consuming. “We’re hoping the faculty will learn enough about the technical operation of their projects that we can devote more staff time to digging deeper,” he says.
Master class lectures by experts who combine humanities and tech
Beginning in 2020, the Program in Digital Humanities will reach out to the wider community — at MIT and in Cambridge and Boston. The plan, Cuthbert says, is to develop a lecture series based on the master class model. Outside experts who combine technology and the humanities in their profession will come to the lab to work with students and then give a public lecture.
The overall goal, Cuthbert says, is to meet a target set by Melissa Nobles, the Kenan Sahin Dean School of Humanities, Arts, and Social Sciences: “to connect the great things going on in computation with the amazing things happening in MIT’s humanities, arts, and social science fields.”
“We have an opportunity to create a love for humanities and an acknowledgement of the importance of humanistic research with the next generation of computer programmers,” Cuthbert says. “We are incredibly excited.”Story prepared by MIT SHASS Communications
Editorial and Design Director: Emily Hiestand
Senior Writer: Kathryn O'Neill
MIT International Science and Technology Initiatives (MISTI) Global Seed Funds (GSF) empower MIT researchers to build collaborations that tackle international problems. Funded projects unite teams of faculty and students with international peers, combining their individual strengths to address challenging issues that may have a worldwide impact.
"This international collaboration was a great encouragement to us,” says GSF awardee Yogesh Surendranath. “The current energy crisis is a global problem, so having scientists from all over the world with different backgrounds, needs, cultures, and expertise come together to address this problem at the chemical level was deeply inspiring."
This year, the 26 funds that comprise MISTI GSF received 221 applications. Over $2 million was awarded to 106 winners from 24 departments across all five schools. That brings the total amount of funding awarded to $17.7 million over the 11-year life of the program.
GSF projects have often culminated in research associated with published papers, and at times have been able to leverage those results to obtain additional research funding. Over two-thirds of the projects report at least one publication that directly resulted from the collaboration, with others referencing a publication in process.
“This worked out better than I could have ever imagined,” says GSF awardee Michael Short. “[My international collaborator] has been conducting research in my area for 25 years, yet never really published in English literature. Now we have three papers together in the last 18 months, and are preparing six more with joint results.”
GSF awardee Vincent Chan also appreciates his impactful partnership.
"As a result of our collaboration in 2017, we successfully created a new program on IoT Smart City funded by the Innovative Technology Commission of Hong Kong with matching funds from the industry” he says. “Also, HKU landed a new program on environmental sensing … and MIT is the international adviser of the project."
GSF projects contribute to the Institute’s educational mission as well. Nearly 75 percent of GSF teams include students from both the MIT and international teams. The next call for proposals will begin in late May.
MIT International Science and Technology Initiatives creates applied international learning opportunities for MIT students that increase their ability to understand and address real-world problems. MISTI collaborates with partners at MIT and beyond, serving as a vital nexus of international activity and bolstering the Institute’s research mission by promoting collaborations between MIT faculty members and their counterparts abroad.
MISTI is a program in the Center for International Studies within the School of Humanities, Arts, and Social Sciences.
In 1988, the U.S. federal government created a $3 billion, 15-year project to sequence the human genome. Not only did the project advance science, it hit the economic jackpot: In 2012, human genome sequencing accounted for an estimated 280,000 jobs, $19 billion in personal income, $3.9 billion in federal taxes, and $2.1 billion in state and local taxes. And all for a price of $2 per year per U.S. resident.
“It’s an incredible rate of return,” says MIT economist Simon Johnson.
It’s not just genomics that pays off. Every additional $10 million in public funding granted to the National Institutes of Health, according to one MIT study, on average produces 2.7 patents and an additional $30 million in value for the private-sector firms that own those patents. When it comes to military technology, each dollar in publicly funded R&D leads to another $2.50-$5.90 in private-sector investment.
In general, “Public investment in science has very big economic returns,” says Johnson, who is the Ronald A. Kurtz Professor of Entrepreneurship at the MIT Sloan School of Management.
Yet after a surge in science funding spurred by World War II, the U.S. has lowered its relative level of public investment in research and development — from about 2 percent of GDP in 1964 to under half of that today.
Reviving U.S. support of science and technology is one of the best ways we can generate economic growth, according to Johnson and his MIT economist colleague Jonathan Gruber, who is the Ford Professor of Economics in MIT’s Department of Economics. And now Johnson and Gruber make that case in a new book, “Jump-Starting America: How Breakthrough Science Can Revive Economic Growth and the American Dream,” published this month by PublicAffairs press.
In it, the two scholars contend that pumping up public investment in science would create not only overall growth but also better jobs throughout the economy, in an era when stagnating incomes have caused strain for a large swath of Americans.
“Good jobs are for MIT graduates, but they’re also for people who don’t finish college. They’re for people who drop out of high school,” says Johnson. “There’s a tremendous amount of anxiety across the country.”
Indeed, spurring growth across the country is a key theme of “Jump-Starting America.” Technology-based growth in the U.S. has been focused in a few “superstar” cities, where high-end tech jobs have been accompanied by increased congestion and sky-high housing prices, forcing out the less well-off.
“The prosperity has been concentrated in some places where it’s become incredibly expensive to live and work,” Johnson says. That includes Silicon Valley, San Francisco, New York, Los Angeles, Seattle, the Washington area, and the Boston metro area.
And yet, Johnson and Gruber believe, the U.S. has scores of cities where the presence of universities combined with industrial know-how could produce more technology-based growth. Some already have: As the authors discuss in the book, Orlando, Florida, is a center for high-tech computer modeling and simulation, thanks to the convergence of federal investment, the growth of the University of Central Florida, and local backing of an adjacent research park that supports dozens of thriving enterprises.
The Orlando case is “a modern version of what once made America the most prosperous nation on Earth,” the authors write, and they believe it can be replicated widely.
“Let’s spread it around the country, to take advantage of where the talent is in the U.S., because there’s a lot of talent away from the coastal cities,” Johnson says.
“Jump-Starting America” even contains a list of 102 metropolitan areas the authors think are ripe for investment and growth, thanks to well-educated work forces and affordability, among other factors. At the top of the list are Pittsburgh, Rochester, and three cities in Ohio: Cincinnati, Columbus, and Cleveland.
The authors’ list does not include any California cities — where affordability is generally a problem — but they view the ranking as a conversation-starter, not the last word on the subject. The book’s website has an interactive feature where readers can tweak the criteria used to rank cities, and see the results.
“We’d like people to challenge us and say, maybe we should think of the criteria differently,” Johnson says. “Everyone should be thinking about what have we got in our region, what do we need to get, and what kind of investment would make the difference here.”
A dividend on your investment
“Jump-Starting America” has received praise from scholars and policy experts. Susan Athey, an economist at Stanford University, calls the book “brilliant” and says it “brings together economic history, urban economics, and the design of incentives to build an ambitious proposal” for growth. Jean Tirole, of the Toulouse School of Economics, says the book gives a boost to industrial policy, by showing “how the government can promote innovation while avoiding the classic pitfalls” of such interventions.
For their part, Johnson and Gruber readily acknowledge that public investment in R&D is just one component of long-term growth. Continued private-sector investment, they note, is vital as well. Still, the book does devote a chapter to the limits of private investment, including the short-term focus on returns that has led many firms to scale back their own R&D operations.
“We’re very pro-private sector,” Johnson says. “I’m a professor of entrepreneurship at Sloan, and I work a lot with entrepreneurs around the world and venture capitalists. They will tell you, quite frankly … their incentives are to make money relatively quickly, given their time horizons and what their investors want. As a result they are drawn to a few sectors, including information technology, and within that more software than hardware these days.”
As a sweetener for any program of public science investment, the authors also suggest that people should receive a kind of annual innovation dividend — a return on their tax dollars. In effect, this would be a scaled-up version of the dividend that, for instance, Alaskans receive on that state’s energy revenues.
That would be a departure from current U.S. policy, but ultimately, Johnson and Gruber say, a departure is what we need.
“We don’t find the existing policies from anyone compelling,” Johnson says. “So we wanted to put some ideas out there and really start a debate about those alternatives, including a return to a bigger investment in science and technology.”
In February, the Institute established five working groups to generate ideas for different components of the structure and operation of the new MIT Stephen A. Schwarzman College of Computing. The Organizational Structure working group is charged with recommending ways to organize the college’s departments and programs, establish its governance, and link it academically with MIT’s five schools, among other considerations. To get a glimpse at the group’s goals and progress, MIT News recently spoke with co-chairs Nelson Repenning, the associate dean of leadership and special projects and the Distinguished Professor of System Dynamics and Organization Studies at the MIT Sloan School of Management; and Asu Ozdaglar, the School of Engineering Distinguished Professor of Engineering and head of the Department of Electrical Engineering and Computer Science (EECS). Three community forums for all five working groups are being held Wednesday, April 17, and Thursday, April 18.
Q: Distilled, what are the key goals of your working group?
Repenning: First, we want to continue to have world-class research and teaching in computer science. Second, we want to develop a structure that will support people outside the field of computer science who are using and developing new computational innovations to do research. Computing is going to be much more tightly connected with the rest of the activity that happens at MIT, so we have to continue to infuse and support computing in all places where it’s relevant.
One challenge we face in developing such a structure is that right now, though they share a department, electrical engineering and computer science are somewhat divided. Our committee feels that categorizing everyone as either/or has outlived its usefulness and is creating needless friction. We aren’t quite sure yet how to resolve this issue, but we are working on it.
Ozdaglar: We are also discussing how to build bi-directional bridges to support interdisciplinary research between computing and other academic disciplines. Another important goal is to incorporate social science not as an afterthought, but as a critical component of future computing research.
Q: What major lessons have you learned in this process?
Repenning: Many of the fields we have at MIT move pretty fast, computer science being the most significant example now. Our organizational structure for teaching and hiring side doesn’t move fast enough to keep up. We are trying to design and reorganize a structure that can adapt to this changing landscape. It’s been challenging considering everyone’s differing views. It’s high stakes. Researchers care a lot about what they do, and don’t want their lives disrupted.
What I didn’t expect when I jumped into this is that we have a chance to set a standard here that could have impact far beyond the [MIT Schwarzman College of Computing]. There have been tectonic shifts in what we do here and, whether it’s genomics or sustainability, we may launch a new entity around another hot field in the near future. So, there will have to be a lot of thinking about how we add more dynamism to our existing organizational structure, so we’re continue to hire the right faculty and prepare students for challenges they’ll face when they graduate.
Q: Where are you now in the process and what’s left to do before submitting your summary report?
Ozdaglar: These new entities are way too complicated to try to design from a clean sheet of paper. We spent quite a bit of time going through a variety of structures we have for organizational research and technology here and elsewhere, and did a thorough diagnosis about what we liked and didn’t like about those structures. We went through EECS, CSAIL, IDSS, ORC, and IMES, and talked about the structures of entities at other schools. Our goal is to evaluate the strength and weaknesses of a bunch of different design options.
Repenning: Now, we have a pretty good sense of the landscape and are evaluating some strong designs. I will say the commitment of everyone at MIT to make this work has been remarkable. We’ve been meeting each week and everyone shows up, even people from departments that may not connect much with the new college. I think it’s a good testament to MIT’s culture.
Four MIT faculty members are among more than 200 leaders from academia, business, public affairs, the humanities, and the arts elected to the American Academy of Arts and Sciences, the academy announced today.
One of the nation’s most prestigious honorary societies, the academy is also a leading center for independent policy research. Members contribute to academy publications, as well as studies of science and technology policy, energy and global security, social policy and American institutions, the humanities and culture, and education.
Those elected from MIT this year are:
- Dimitri A. Antoniadis, Ray and Maria Stata Professor of Electrical Engineering;
- Anantha P. Chandrakasan, dean of the School of Engineering and the Vannevar Bush Professor of Electrical Engineering and Computer Science;
- Guoping Feng, the James W. (1963) and Patricia T. Poitras Professor of Brain and Cognitive Sciences; and
- David R. Karger, professor of electrical engineering.
“We are pleased to recognize the excellence of our new members, celebrate their compelling accomplishments, and invite them to join the academy and contribute to its work,” said David W. Oxtoby, president of the American Academy of Arts and Sciences. “With the election of these members, the academy upholds the ideals of research and scholarship, creativity and imagination, intellectual exchange and civil discourse, and the relentless pursuit of knowledge in all its forms.”
The new class will be inducted at a ceremony in October in Cambridge, Massachusetts.
Since its founding in 1780, the academy has elected leading “thinkers and doers” from each generation, including George Washington and Benjamin Franklin in the 18th century, Maria Mitchell and Daniel Webster in the 19th century, and Toni Morrison and Albert Einstein in the 20th century. The current membership includes more than 200 Nobel laureates and 100 Pulitzer Prize winners.
A new learning system developed by MIT researchers improves robots’ abilities to mold materials into target shapes and make predictions about interacting with solid objects and liquids. The system, known as a learning-based particle simulator, could give industrial robots a more refined touch — and it may have fun applications in personal robotics, such as modelling clay shapes or rolling sticky rice for sushi.
In robotic planning, physical simulators are models that capture how different materials respond to force. Robots are “trained” using the models, to predict the outcomes of their interactions with objects, such as pushing a solid box or poking deformable clay. But traditional learning-based simulators mainly focus on rigid objects and are unable to handle fluids or softer objects. Some more accurate physics-based simulators can handle diverse materials, but rely heavily on approximation techniques that introduce errors when robots interact with objects in the real world.
In a paper being presented at the International Conference on Learning Representations in May, the researchers describe a new model that learns to capture how small portions of different materials — “particles” — interact when they’re poked and prodded. The model directly learns from data in cases where the underlying physics of the movements are uncertain or unknown. Robots can then use the model as a guide to predict how liquids, as well as rigid and deformable materials, will react to the force of its touch. As the robot handles the objects, the model also helps to further refine the robot’s control.
In experiments, a robotic hand with two fingers, called “RiceGrip,” accurately shaped a deformable foam to a desired configuration — such as a “T” shape — that serves as a proxy for sushi rice. In short, the researchers’ model serves as a type of “intuitive physics” brain that robots can leverage to reconstruct three-dimensional objects somewhat similarly to how humans do.
“Humans have an intuitive physics model in our heads, where we can imagine how an object will behave if we push or squeeze it. Based on this intuitive model, humans can accomplish amazing manipulation tasks that are far beyond the reach of current robots,” says first author Yunzhu Li, a graduate student in the Computer Science and Artificial Intelligence Laboratory (CSAIL). “We want to build this type of intuitive model for robots to enable them to do what humans can do.”
“When children are 5 months old, they already have different expectations for solids and liquids,” adds co-author Jiajun Wu, a CSAIL graduate student. “That’s something we know at an early age, so maybe that’s something we should try to model for robots.”
Joining Li and Wu on the paper are: Russ Tedrake, a CSAIL researcher and a professor in the Department of Electrical Engineering and Computer Science (EECS); Joshua Tenenbaum, a professor in the Department of Brain and Cognitive Sciences and a member of CSAIL and the Center for Brains, Minds, and Machines (CBMM); and Antonio Torralba, a professor in EECS and director of the MIT-IBM Watson AI Lab.
A key innovation behind the model, called the “particle interaction network” (DPI-Nets), was creating dynamic interaction graphs, which consist of thousands of nodes and edges that can capture complex behaviors of so-called particles. In the graphs, each node represents a particle. Neighboring nodes are connected with each other using directed edges, which represent the interaction passing from one particle to the other. In the simulator, particles are hundreds of small spheres combined to make up some liquid or a deformable object.
The graphs are constructed as the basis for a machine-learning system called a graph neural network. In training, the model over time learns how particles in different materials react and reshape. It does so by implicitly calculating various properties for each particle — such as its mass and elasticity — to predict if and where the particle will move in the graph when perturbed.
The model then leverages a “propagation” technique, which instantaneously spreads a signal throughout the graph. The researchers customized the technique for each type of material — rigid, deformable, and liquid — to shoot a signal that predicts particles positions at certain incremental time steps. At each step, it moves and reconnects particles, if needed.
For example, if a solid box is pushed, perturbed particles will be moved forward. Because all particles inside the box are rigidly connected with each other, every other particle in the object moves the same calculated distance, rotation, and any other dimension. Particle connections remain intact and the box moves as a single unit. But if an area of deformable foam is indented, the effect will be different. Perturbed particles move forward a lot, surrounding particles move forward only slightly, and particles farther away won’t move at all. With liquids being sloshed around in a cup, particles may completely jump from one end of the graph to the other. The graph must learn to predict where and how much all affected particles move, which is computationally complex.
Shaping and adapting
In their paper, the researchers demonstrate the model by tasking the two-fingered RiceGrip robot with clamping target shapes out of deformable foam. The robot first uses a depth-sensing camera and object-recognition techniques to identify the foam. The researchers randomly select particles inside the perceived shape to initialize the position of the particles. Then, the model adds edges between particles and reconstructs the foam into a dynamic graph customized for deformable materials.
Because of the learned simulations, the robot already has a good idea of how each touch, given a certain amount of force, will affect each of the particles in the graph. As the robot starts indenting the foam, it iteratively matches the real-world position of the particles to the targeted position of the particles. Whenever the particles don’t align, it sends an error signal to the model. That signal tweaks the model to better match the real-world physics of the material.
Next, the researchers aim to improve the model to help robots better predict interactions with partially observable scenarios, such as knowing how a pile of boxes will move when pushed, even if only the boxes at the surface are visible and most of the other boxes are hidden.
The researchers are also exploring ways to combine the model with an end-to-end perception module by operating directly on images. This will be a joint project with Dan Yamins’s group; Yamin recently completed his postdoc at MIT and is now an assistant professor at Stanford University. “You’re dealing with these cases all the time where there’s only partial information,” Wu says. “We’re extending our model to learn the dynamics of all particles, while only seeing a small portion.”
The MIT D-Lab Scale-Ups fellowship program, which offers one year of support to social entrepreneurs bringing poverty-alleviating products to market at scale, has announced its six fellows for 2019. This year’s fellows include the founders of homegrown, high-impact ventures in underserved markets in Kenya, Tanzania, and Uganda.
“I look forward to learning from — and sharing with — other D-Lab Scale-Ups Fellows and the team at MIT, and to establishing a strong and reliable network to accelerate our growth,” says Peter Nyamai, founder of the Global Expressions Group in Kenya and a new fellow in the program.
All six founders are native to and currently living in their markets.
“We are excited to work with a vibrant cohort of East African entrepreneurs whose expertise is grounded in their lived reality,” says Jona Repishti, who manages the fellowship. “Working with local founders has certain advantages — they reflect the demographics of the markets they serve; their lived experience helps them identify unique, scalable, market-based solutions overlooked by outsiders. What’s more, they are more likely to commit for the long haul, developing local talent and infrastructure along the way.”
During the yearlong program, the D-Lab Scale-Ups Fellows work to retire risk and position their ventures for investment, partnership, and growth. Each social entrepreneur receives a $20,000 grant, tailored mentoring, skills building workshops, and networking opportunities. They also participate in a fellows’ retreat, this year taking place at D-Lab in early May.
“As individuals and as a cohort, these fellows have great change potential for their regional ecosystems,” Repishti notes. “And by bringing them into D-Lab and the broader MIT community, we hope to advance not only their ventures, but also the D-Lab and MIT approach to social entrepreneurship.”
Launched in 2012, the D-Lab Scale-Ups fellowship has supported 39 fellows working on four continents in sectors including agriculture, energy, water, health care, housing, mobility, recycling, education, and personal finance. At the close of last year’s cycle, Scale-Ups fellows had raised $11.1 million in funding, generated $10.2 million in revenue, created over 700 direct and 6,700 indirect full-time equivalent jobs, and reached 1.5 million people living in low-income settings with their product and service offerings.
“Over the course of the 12-month fellowship we tackle the knowledge gap by putting fellows in the driver’s seat,” Repishti explains. “Our entrepreneurs guide us to provide support that is tailored to address their pain points, delivered fast, and focused on the essentials. Our curriculum is experiential, dynamic, and focused on transforming mindsets, just capabilities.”
This year’s D-Lab Scale-Ups Fellows are:
Winnie Gitau (Kenya)
Winnie Gitau is founder of Kwangu Kwako, which provides safer, healthier, and more secure housing alternatives to the traditional informal settlement structures in Kenya. To do this, they use reinforced pre-cast concrete panels made by local artisans within the community. The outcome is a safer, simpler and more cost-effective alternative to the existing Mabati structures, which are temporary homes built from corrugated iron sheets and bush poles.
This summer, as part of her fellowship, she says she will have the opportunity to work with a D-Lab monitoring and evaluation fellow to figure out “little measurable ways to gauge” the human impact of Kwangu Kwako’s work, in other words: “how to best measure our impact besides the number of houses or units sold.”
Dysmus Kisilu (Kenya)
Dysmus Kisilu, winner of a 2018 Gates Foundation Goalkeeper Award, is the founder of Solar Freeze, a Kenya-based enterprise that has pioneered mobile cold storage units powered by renewable energy to help rural smallholder farmers reduce postharvest loss. In many developing countries, postharvest losses are as high as 80% and the cold‐storage chain is virtually non‐existent due to high equipment costs and spotty electricity. Solar Freeze has provided solar-powered irrigation kits and cold storage devices to more than 3,000 farmers.
Christian Mwilage (Tanzania)
“I expect to gain connections, partnerships, knowledge sharing, and training that will help us strengthen and scale our project,” says Christian Mwijage, the founder of EcoAct Tanzania. EocAct is a for‐profit social enterprise that has developed a chemical-free and energy-conserving technology which transforms post-consumer waste plastics into durable plastic timbers for use in construction. EcoAct plastic timber is an affordable alternative to traditional timber, reduces the need for building material manufactured from wood, and helps to preserve forests and mitigate climate change.
Peter Mumo Nyamai (Kenya)
Peter Mumo Nyamai is the founder of Expressions Global Group, a social venture which supplies innovative, durable, and environment-friendly rainwater harvesting products to improve irrigation and boost productivity among rural smallholder farmers in Kenya. In addition to its product line, Expressions Global links farmers to the affordable credit and ready markets for fresh produce. “We are at a point of growth where we needed help in fine-tuning key segments of our venture,” Nyamai says, “and the MIT D-Lab Scale-Ups Fellowship offers not only a cocktail of everything we needed, but a chance to be part of community driven by a shared purpose: to make a better world through education, research, and innovation.”
Chrispinus Onyancha (Uganda)
“Winning this fellowship is a big confirmation,” says Chrispinus Onyancha, CEO of clinicPesa, a platform established to provide access to health care financing to individuals in East Africa to offset medical bills and buy medications at any clinicPesa-registered clinic, hospital, or pharmacy. “We strongly believe that creative health care financing should be among key priority focus areas, especially for marginalized groups in Africa where infant mortality rates are so high and over 20 million people per year till die of preventable and treatable conditions such as malaria.”
Prince Prosper Tillya (Tanzania)
Prince Prosper Tyllya is the founder and managing director of FixChap, a digital platform in Tanzania through which clients can book repair requests and get connected instantly to verified local handymen, sourced from vocational training institutions. “Winning the fellowship is an incredible milestone for us,” Tillya says. “I hope the fellowship will take our venture to a whole new level on every aspect which will enables us to increase our impact and touch more lives.” FixChap benefits clients such as homeowners and businesses by providing reliable and competent services while providing a stable income to the handymen.
It was 11:59 p.m. on Pi Day, March 14, in the MIT Alumni Association (MITAA) headquarters. The lights were still on. As midnight arrived, a cheer went up among the MIT Annual Fund staffers there working late into the night. The third annual MIT 24-Hour Challenge had not only reached its original target of unlocking a $314,159 challenge gift from an anonymous donor, but had inspired participation from enough donors to obtain a $50,000 bonus from a second benefactor and to meet more than 60 microchallenges issued by supporters of programs and groups across MIT.
“One thing is clear,” MIT Alumni Association CEO Whitney T. Espich wrote in a note to all Institute alumni and friends announcing the results. “MIT is fortunate to have this one-of-a-kind community. Your generosity, care, and dedication fuel the future at MIT.”
In total, a record 9,397 donors made gifts to MIT over the course of just one day, totaling $3.75 million. (Anyone interested can learn more about the final tallies on the MIT 24-Hour Challenge website.)
Of course, the team in the MIT 24-Hour Challenge HQ couldn’t have brought in so many gifts on its own. All across the world, MIT alumni and friends were celebrating the unofficial nerd holiday that is 3.14, with many featuring Challenge-related drives. On campus, a Pi Day Fair was set up in Lobby 10, complete with pies, a Pi recitation contest, and a Cash Cube, in which students grabbed for dollars that could be distributed to areas of focus for MIT’s Student Philanthropy Program.
All of these moments led to a Pi Day for the books. The MIT Annual Fund credits the success to the collaboration across the entire Institute and throughout the widespread community of MIT alumni, students, and friends. Among those working hardest on March 14 to activate that network were a cohort of 544 volunteer Web Ambassadors.
Each ambassador received a personalized link to share with those they encouraged to give during the challenge. As Pi Day went on, the ambassadors were able to track the progress of their efforts, which ranged from social media posts to phone calls to emails — many, many emails. In all, 995 gifts were linked to the outreach of Web Ambassadors.
For three volunteers whose efforts met particular success that day — each bringing in between 35 and 50 gifts — the payoff went beyond seeing the dollars add up. Each was working to raise funds for a particular program close to their hearts. Yet, as they tell it, the day also cemented the connection they feel to the MIT community, which has proven again and again that it can join forces to become far more than the sum of its parts.
Those three volunteers shared why they stepped up during the MIT 24-Hour Challenge.
José Zavala, MIT junior and chair of the Lecture Series Committee (LSC):
“LSC is a student-run organization that runs an on-campus movie theater and hosts special lecture series events. We successfully held a microchallenge on Pi Day for a new projection and sound system that will help us stay on top of the future of the digital era.
“The most MIT thing we experienced during the 24-Hour Challenge this year was reconnecting with our alumni. One alumnus replied to us with a success story about a showing of 'Zorba the Greek' he ran all the way back in the '60s. We loved hearing about the days when LSC ran shows in Kresge and sold out houses constantly. What we loved the most about his response is the photo he included of them celebrating in the LSC office afterwards.
“It is beyond words to express how faithful our alumni have been to LSC. They’ve shown up to help us reacquaint ourselves with old equipment and rushed to the scene when we messaged out that we needed assistance urgently. One alum works for the projection company that supplied us with the most affordable quote for the projector for which we just successfully fundraised. Another signed up to be our donation matcher for this big challenge we completed. Others offered consulting, publicity, advice, and stories from their time in LSC. All of them were invaluable in achieving our fundraising goal.
“Thank you, LSC alumni. Your generosity and goodwill go to show the decades-old LSC motto we are all familiar with: It doesn’t suck.”
Charlotte Farquhar, first-year graduate student in chemistry:
“The MIT chemistry department is my home for the next four years, at least. The Chemistry Alliance for Diversity and Inclusion is dedicated to making that department a more diverse and inclusive environment, and providing a space to engage with each other and share the diversity that already exists. The additional funding we raised on Pi Day can go toward improving the experience of everyone in our department, through a mentoring program for underrepresented minority applicants, monthly open forums and events within the department, and professional development opportunities with chemists from around the country.
“The message was simple: Help make MIT a more inclusive environment where everyone can feel comfortable and supported. Most of our donors ended up coming from inside the department, and we were really proud of the chemistry graduate students during the MIT 24-Hour Challenge. Hitting our microchallenge goal of 100 donors on Pi Day was really a team effort.”
Monica Reynolds ’82, vice president of fundraising, MIT Crew Alumni Association Board of Directors:
“It is really easy for me to support MIT and ask people to contribute. It’s easy to sell something you honestly believe in. Why MIT Crew? I rowed crew for all four years. I didn’t think much about the character-building aspects of the sport until much later in life. As I’ve grown up, I’ve learned a lot about how the cold, cruel world works. Money really runs everything and there is only so much to go around. Crew is expensive. Alumni contributions are a big part of keeping it afloat — pun intended.
“Why else do I do this? Fundraising for MIT is something I can squeeze into my schedule wherever it fits — kind of like stuffing the little odds and ends into the corners of your suitcase. It’s also a fun challenge, and it has forced me outside of my comfort zone. Having MIT as a partner makes this all the more fun because I know that they have the know-how and technology to make ideas happen.
“I have been an MIT ambassador for three years. Being the goal-oriented, efficiency-focused, occasionally socially-awkward MIT alum that I am — and as many of us are — I started out my fundraising attempts being very blunt, just admitting up front that I wanted them to donate money. Result: limited success. I had some help from the staff at the Annual Fund and some subtle suggestions for softening the message. I also learned the magic of thank you for something as simple as opening the email. Then you have to encourage them to think about why they might want to give. What about MIT makes it a positive experience in their life? And finally, I learned that you need to speak with your own voice. Be sincere and grateful.
“Throughout the entire Pi Day effort, building up to it and during the day, everyone was pulling together — another crew pun, sorry. Collaboration rather than cutthroat competition is what I felt when I was a student at MIT, and it is still what I feel when I work with MIT alum and staff.
“One exception, and I didn’t want to ever admit this: Steve Larky — watch your back! I got up just before midnight to log in the first crew donation, but you still beat me to it!”
Every year trash companies sift through an estimated 68 million tons of recycling, which is the weight equivalent of more than 30 million cars.
A key step in the process happens on fast-moving conveyor belts, where workers have to sort items into categories like paper, plastic and glass. Such jobs are dull, dirty, and often unsafe, especially in facilities where workers also have to remove normal trash from the mix.
With that in mind, a team led by researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) has developed a robotic system that can detect if an object is paper, metal, or plastic.
The team’s “RoCycle” system includes a soft Teflon hand that uses tactile sensors on its fingertips to detect an object’s size and stiffness. Compatible with any robotic arm, RoCycle was found to be 85 percent accurate at detecting materials when stationary, and 63 percent accurate on an actual simulated conveyer belt. (Its most common error was identifying paper-covered metal tins as paper, which the team says would be improved by adding more sensors along the contact surface.)
“Our robot’s sensorized skin provides haptic feedback that allows it to differentiate between a wide range of objects, from the rigid to the squishy,” says MIT Professor Daniela Rus, senior author on a related paper that will be presented in April at the IEEE International Conference on Soft Robotics (RoboSoft) in Seoul, South Korea. “Computer vision alone will not be able to solve the problem of giving machines human-like perception, so being able to use tactile input is of vital importance.”
A collaboration with Yale University, RoCycle directly demonstrates the limits of sight-based sorting: It can reliably distinguish between two visually similar Starbucks cups, one made of paper and one made of plastic, that would give vision systems trouble.
Rus says that the project is part of her larger goal to reduce the back-end cost of recycling, in order to incentivize more cities and countries to create their own programs. Today recycling centers aren’t particularly automated; their main kinds of machinery include optical sorters that use different wavelength light to distinguish between plastics, magnetic sorters that separate out iron and steel products, and aluminum sorters that use eddy currents to remove non-magnetic metals.
This is a problem for one very big reason: just last month China raised its standards for the cleanliness of recycled goods it accepts from the United States, meaning that some of the country’s single-stream recycling is now sent to landfills.
"If a system like RoCycle could be deployed on a wide scale, we'd potentially be able to have the convenience of single-stream recycling with the lower contamination rates of multi-stream recycling,” says PhD student Lillian Chin, lead author on the new paper.
It’s surprisingly hard to develop machines that can distinguish between paper, plastic, and metal, which shows how impressive a feat it is for humans. When we pick up an object, we can immediately recognize many of its qualities even with our eyes closed, like whether it’s large and stiff or small and soft. By feeling the object and understanding how that relates to the softness of our own fingertips, we are able to learn how to handle a wide range of objects without dropping or breaking them.
This kind of intuition is tough to program into robots. Traditional hard (“rigid”) robot hands have to know an object’s exact location and size to be able to calculate a precise motion path. Soft hands made of materials like rubber are much more flexible, but have a different problem: Because they’re powered by fluidic forces, they have a balloon-like structure that can puncture quite easily.
How RoCycle works
Rus’ team used a motor-driven hand made of a relatively new material called “auxetics.” Most materials get narrower when pulled on, like a rubber band when you stretch it; auxetics, meanwhile, actually get wider. The MIT team took this concept and put a twist on it, quite literally: They created auxetics that, when cut, twist to either the left or right. Combining a “left-handed” and “right-handed” auxetic for each of the hand’s two large fingers makes them interlock and oppose each other’s rotation, enabling more dynamic movement. (The team calls this “handed-shearing auxetics”, or HSA.)
“In contrast to soft robots, whose fluid-driven approach requires air pumps and compressors, HSA combines twisting with extension, meaning that you’re able to use regular motors,” says Chin.
The team’s gripper first uses its “strain sensor” to estimate an object’s size, and then uses its two pressure sensors to measure the force needed to grasp an object. These metrics — along with calibration data on the size and stiffnesses of objects of different material types — are what gives the gripper a sense of what material the object is made. (Since the tactile sensors are also conductive, they can detect metal by how much it changes the electrical signal.)
“In other words, we estimate the size and measure the pressure difference between the current closed hand and what a normal open hand should look like,” says Chin. “We use this pressure difference and size to classify the specific object based on information about different objects that we’ve already measured.”
RoCycle builds on an set of sensors that detect the radius of an object to within 30 percent accuracy, and tell the difference between “hard” and “soft” objects with 78 percent accuracy. The team’s hand is also almost completely puncture resistant: It was able to be scraped by a sharp lid and punctured by a needle more than 20 times, with minimal structural damage.
As a next step, the researchers plan to build out the system so that it can combine tactile data with actual video data from a robot’s cameras. This would allow the team to further improve its accuracy and potentially allow for even more nuanced differentiation between different kinds of materials.
Chin and Rus co-wrote the RoCycle paper alongside MIT postdoc Jeffrey Lipton, as well as PhD student Michelle Yuen and Professor Rebecca Kramer-Bottiglio of Yale University.
This project was supported in part by Amazon, JD.com, the Toyota Research Institute, and the National Science Foundation.
Computational neuroscientist Sarah Schwettmann is one of three instructors behind the cross-disciplinary course 9.S52/9.S916 (Vision in Art and Neuroscience), which introduces students to core concepts in visual perception through the lenses of art and neuroscience. Supported by a faculty grant from the Center for Art, Science and Technology at MIT (CAST) for the past two years, the class is led by Pawan Sinha, a professor of vision and computational neuroscience in the Department of Brain and Cognitive Sciences. They are joined in the course by Seth Riskin SM ’89, a light artist and the manager of the MIT Museum Studio and Compton Gallery, where the course is taught. Schwettman discussed the combination of art and science in an educational setting.
Q: How have the three of you approached this cross-disciplinary class in art and neuroscience?
A: Discussions around this intersection often consider what each field has to offer the other. We take a different approach, one I refer to as occupying the gap, or positioning ourselves between the two fields and asking what essential questions underlie them both. One question addresses the nature of the human relationship to the world. The course suggests one answer: This relationship is fundamentally creative, from the brain’s interpretation of incoming sensory data in perception, to the explicit construction of experiential worlds in art.
Neuroscience and art, therefore, each provide a set of tools for investigating different levels of the constructive process. Through neuroscience, we develop a specific understanding of the models of the world that the brain uses to make sense of incoming visual data. With articulation of those models, we can engineer types of inputs that interact with visual processing architecture in particularly exquisite ways, and do so reliably, giving artists a toolkit for remixing and modulating experience. In the studio component of the course, we experiment with this toolkit and collectively move it forward.
While designing the course, Pawan, Seth, and I found that we were each addressing a similar set of questions, the same that motivate the class, through our own research and practice. In parallel to computational vision research, Professor Sinha leads a humanitarian initiative called Project Prakash, which provides treatment to blind children in India and explores the development of vision following the restoration of sight. Where does structure in perception originate? As an artist in the MIT Museum Studio, Seth works with articulated light to sculpt structured visual worlds out of darkness. I also live on this interface where the brain meets the world — my research in the Department of Brain and Cognitive Sciences examines the neural basis of mental models for simulating physics. Linking our work in the course is an experiment in synthesis.
Q: What current research in vision, neuroscience, and art are being explored at MIT, and how does the class connect it to hands-on practice?
A: Our brains build a rich world of experience and expectation from limited and noisy sensory data with infinite potential interpretations. In perception research, we seek to discover how the brain finds more meaning in incoming data than is explained by the signal alone. Work being done at MIT around generative models addresses this, for instance in the labs of Josh Tenenbaum and Josh McDermott in the Department of Brain and Cognitive Sciences. Researchers present an ambiguous visual or auditory stimulus and by probing someone’s perceptual interpretation, they get a handle on the structures that the mind generates to interpret incoming data, and they can begin to build computational models of the process.
In Vision in Art and Neuroscience, we focus on the experiential as well as the experimental, probing the perceiver’s experience of structure-generating process—perceiving perception itself. As instructors, we face the pedagogical question: what exercises, in the studio, can evoke so striking an experience of students’ own perception that cutting edge research takes on new meaning, understood in the immediacy of seeing? Later in the semester, students face a similar question as artists: How can one create visual environments where viewers experience their own perceptual processing at work? Done well, this experience becomes the artwork itself. Early in the course, students explore the Ganzfeld effect, popularized by artist James Turrell, where the viewer is exposed to an unstructured visual field of uniform illumination. In this experience, one feels the mind struggling to fit models of the world to unstructured input, and attempting this over and over again — an interpretation process which often goes unnoticed when input structure is expected by visual processing architecture. The progression of the course modules follows the hierarchy of visual processing in the brain, which builds increasingly complex interpretations of visual inputs, from brightness and edges to depth, color, and recognizable form.
MIT students first encounter those concepts in the seminar component of the course at the beginning of each week. Later in the week, students translate findings into experimental approaches in the studio. We work with light directly, from introducing a single pinpoint of light into an otherwise completely dark room, to building intricate environments using programmable electronics. Students begin to take this work into their own hands, in small groups and individually, culminating in final projects for exhibition. These exhibitions are truly a highlight of the course. They’re often one of the first times that students have built and shown artworks. That’s been a gift to share with the broader MIT community, and a great learning experience for students and instructors alike.
Q: How has that approach been received by the MIT community?
A: What we’re doing has resonated across disciplines: In addition to neuroscience, we have students and researchers joining us from computer science, mechanical engineering, mathematics, the Media Lab, and ACT [the Program in Art, Culture, and Technology]. The course is growing into something larger, a community of practice interested in applying the scientific methodology we develop to study the world, to probe experience, and to articulate models for its generation and replication.
With a mix of undergraduates, graduates, faculty, and artists, we’ve put together installations and symposia — including three on campus so far. The first of these, "Perceiving Perception," also led to a weekly open studio night where students and collaborators convene for project work. Our second exhibition, "Dessert of the Real," is on display this spring in the Compton Gallery. This April we’re organizing a symposium in the studio featuring neuroscientists, computer scientists, artists and researchers from MIT and Harvard. We’re reaching beyond campus as well, through off-site installations, collaborations with museums — including the Metropolitan Museum of Art and the Peabody Essex Museum — and a partnership with the ZERO Group in Germany.
We’re eager to involve a broad network of collaborators. It’s an exciting moment in the fields of neuroscience and computing; there is great energy to build technologies that perceive the world like humans do. We stress on the first day of class that perception is a fundamentally creative act. We see the potential for models of perception to themselves be tools for scaling and translating creativity across domains, and for building a deeply creative relationship to our environment.
NASA’s Transiting Exoplanet Survey Satellite, TESS, has discovered its first Earth-sized exoplanet. The planet, named HD 21749c, is the smallest world outside our solar system that TESS has identified yet.
In a paper published today in the journal Astrophysical Journal Letters, an MIT-led team of astronomers reports that the new planet orbits the star HD 21749 — a very nearby star, just 52 light years from Earth. The star also hosts a second planet — HD 21749b — a warm “sub-Neptune” with a longer, 36-day orbit, which the team reported previously and now details further in the current paper.
The new Earth-sized planet is likely a rocky though uninhabitable world, as it circles its star in just 7.8 days — a relatively tight orbit that would generate surface temperatures on the planet of up to 800 degrees Fahrenheit.
The discovery of this Earth-sized world is nevertheless exciting, as it demonstrates TESS’ ability to pick out small planets around nearby stars. In the near future, the TESS team expects the probe should reveal even colder planets, with conditions more suitable for hosting life.
“For stars that are very close by and very bright, we expected to find up to a couple dozen Earth-sized planets,” says lead author and TESS member Diana Dragomir, a postdoc in MIT’s Kavli Institute for Astrophysics and Space Research. “And here we are — this would be our first one, and it’s a milestone for TESS. It sets the path for finding smaller planets around even smaller stars, and those planets may potentially be habitable.”
TESS has been hunting for planets beyond our solar system since it launched on April 18, 2018. The satellite is a NASA Astrophysics Explorer mission that is led and operated by MIT, and is designed to observe nearly the entire sky, in overlapping, month-long patches, or “sectors,” as it orbits the Earth. As it circles our own planet, TESS focuses its four cameras outward to monitor the nearest, brightest stars in the sky, looking for any periodic dips in starlight that could indicate the presence of an exoplanet as it passes in front of its host star.
Over its two-year mission, TESS aims to identify for the astronomy community at least 50 small, rocky planets, along with estimates of their masses. To date, the mission has discovered 10 planets smaller than Neptune, four of their masses which have been estimated, including π Men b, a planet twice the size of Earth, with a six-day orbit around its star; LHS 3844b, a hot, rocky world that’s slightly bigger than Earth and circles its star in a blistering 11 hours; and TOI 125b and c — two “sub-Neptunes” that orbit the same star, both within about a week. All four of these planets were identified from data obtained during TESS’ first two observing sectors — a good indication, the team writes in its paper, that “many more are to be found.”
Dragomir picked out this newest, Earth-sized planet from the first four sectors of TESS observations. When these data became available, in the form of light curves, or intensities of starlight, she fed them into a software code to look for interesting, periodic signals. The code first identified a possible transit that the team later confirmed as the warm sub-Neptune they announced earlier this year.
As is usually the case with small planets, where there’s one, there are likely to be more, and Dragomir and her colleagues decided to comb through the same observations again to see if they could spot any other small worlds hiding in the data.
“We know these planets often come in families,” Dragomir says. “So we searched all the data again, and this small signal came up.”
The team identified a small dip in the light from HD 21749, that occurred every 7.8 days. Ultimately, the researchers identified 11 such periodic dips, or transits, and determined that the star’s light was being momentarily blocked by a planet about the size of the Earth.
While this is the first Earth-sized planet discovered by TESS, other Earth-sized exoplanets have been discovered in the past, mainly by NASA’s Kepler Space Telescope, a since-retired telescope that monitored more than 530,000 stars. In the end, the Kepler mission detected 2,662 planets, many of which were Earth-sized, and a handful of those were deemed to be within their star’s habitable zone — where a balance of conditions could be suitable for hosting life.
However, Kepler observed stars that are many leagues further away than those that are monitored by TESS. Therefore, Dragomir says that following up on any of Kepler’s far-flung, Earth-sized planets would be much harder than studying planets orbiting TESS’ much closer, brighter stars.
“Because TESS monitors stars that are much closer and brighter, we can measure the mass of this planet in the very near future, whereas for Kepler’s Earth-sized planets, that was out of the question,” Dragomir says. “So this new TESS discovery could lead to the first mass measurement of an Earth-sized planet. And we’re excited about what that mass could be. Will it be Earth’s mass? Or heavier? We don’t really know.”
Before being tested in animals or humans, most cancer drugs are evaluated in tumor cells grown in a lab dish. However, in recent years, there has been a growing realization that the environment in which these cells are grown does not accurately mimic the natural environment of a tumor, and that this discrepancy could produce inaccurate results.
In a new study, MIT biologists analyzed the composition of the interstitial fluid that normally surrounds pancreatic tumors, and found that its nutrient composition is different from that of the culture medium normally used to grow cancer cells. It also differs from blood, which feeds the interstitial fluid and removes waste products.
The findings suggest that growing cancer cells in a culture medium more similar to this fluid could help researchers better predict how experimental drugs will affect cancer cells, says Matthew Vander Heiden, an associate professor of biology at MIT and a member of the Koch Institute for Integrative Cancer Research.
“It’s kind of an obvious statement that the tumor environment is important, but I think in cancer research the pendulum had swung so far toward genes, people tended to forget that,” says Vander Heiden, one of the senior authors of the study.
Alex Muir, a former Koch Institute postdoc who is now an assistant professor at the University of Chicago, is also a senior author of the paper, which appears in the April 16 edition of the journal eLife. The lead author of the study is Mark Sullivan, an MIT graduate student.
Scientists have long known that cancer cells metabolize nutrients differently than most other cells. This alternative strategy helps them to generate the building blocks they need to continue growing and dividing, forming new cancer cells. In recent years, scientists have sought to develop drugs that interfere with these metabolic processes, and one such drug was approved to treat leukemia in 2017.
An important step in developing such drugs is to test them in cancer cells grown in a lab dish. The growth medium typically used to grow these cells includes carbon sources (such as glucose), nitrogen, and other nutrients. However, in the past few years, Vander Heiden’s lab has found that cancer cells grown in this medium respond differently to drugs than they do in mouse models of cancer.
David Sabatini, a member of the Whitehead Institute and professor of biology at MIT, has also found that drugs affect cancer cells differently if they are grown in a medium that resembles the nutrient composition of human plasma, instead of the traditional growth medium.
“That work, and similar results from a couple of other groups around the world, suggested that environment matters a lot,” Vander Heiden says. “It really was a wake up call for us that to really know how to find the dependencies of cancer, we have to get the environment right.”
To that end, the MIT team decided to investigate the composition of interstitial fluid, which bathes the tissue and carries nutrients that diffuse from blood flowing through the capillaries. Its composition is not identical to that of blood, and in tumors, it can be very different because tumors often have poor connections to the blood supply.
The researchers chose to focus on pancreatic cancer in part because it is known to be particularly nutrient-deprived. After isolating interstitial fluid from pancreatic tumors in mice, the researchers used mass spectrometry to measure the concentrations of more than 100 different nutrients, and discovered that the composition of the interstitial fluid is different from that of blood (and from that of the culture medium normally used to grow cells). Several of the nutrients that the researchers found to be depleted in tumor interstitial fluid are amino acids that are important for immune cell function, including arginine, tryptophan, and cystine.
Not all nutrients were depleted in the interstitial fluid — some were more plentiful, including the amino acids glycine and glutamate, which are known to be produced by some cancer cells.
Location, location, location
The researchers also compared tumors growing in the pancreas and the lungs and found that the composition of the interstitial fluid can vary based on tumors’ location in the body and at the site where the tumor originated. They also found slight differences between the fluid surrounding tumors that grew in the same location but had different genetic makeup; however, the genetic factors tested did not have as big an impact as the tumor location.
“That probably says that what determines what nutrients are in the environment is heavily driven by interactions between cancer cells and noncancer cells within the tumor,” Vander Heiden says.
Scientists have previously discovered that those noncancer cells, including supportive stromal cells and immune cells, can be recruited by cancer cells to help remake the environment around the tumor to promote cancer survival and spread.
Vander Heiden’s lab and other research groups are now working on developing a culture medium that would more closely mimic the composition of tumor interstitial fluid, so they can explore whether tumor cells grown in this environment could be used to generate more accurate predictions of how cancer drugs will affect cells in the body.
The research was funded by the National Institutes of Health, the Lustgarten Foundation, the MIT Center for Precision Cancer Medicine, Stand Up to Cancer, the Howard Hughes Medical Institute, and the Ludwig Center at MIT.