Transistors, the building blocks of modern electronics, are typically made of silicon. Because it’s a semiconductor, this material can control the flow of electricity in a circuit. But silicon has fundamental physical limits that restrict how compact and energy-efficient a transistor can be.

MIT researchers have now replaced silicon with a magnetic semiconductor, creating a magnetic transistor that could enable smaller, faster, and more energy-efficient circuits. The material’s magnetism strongly influences its electronic behavior, leading to more efficient control of the flow of electricity. 

The team used a novel magnetic material and an optimization process that reduces the material’s defects, which boosts the transistor’s performance.

The material’s unique magnetic properties also allow for transistors with built-in memory, which would simplify circuit design and unlock new applications for high-performance electronics.

“People have known about magnets for thousands of years, but there are very limited ways to incorporate magnetism into electronics. We have shown a new way to efficiently utilize magnetism that opens up a lot of possibilities for future applications and research,” says Chung-Tao Chou, an MIT graduate student in the departments of Electrical Engineering and Computer Science (EECS) and Physics, and co-lead author of a paper on this advance.

Chou is joined on the paper by co-lead author Eugene Park, a graduate student in the Department of Materials Science and Engineering (DMSE); Julian Klein, a DMSE research scientist; Josep Ingla-Aynes, a postdoc in the MIT Plasma Science and Fusion Center; Jagadeesh S. Moodera, a senior research scientist in the Department of Physics; and senior authors Frances Ross, TDK Professor in DMSE; and Luqiao Liu, an associate professor in EECS, and a member of the Research Laboratory of Electronics; as well as others at the University of Chemistry and Technology in Prague. The paper appears today in Physical Review Letters.

Overcoming the limits

In an electronic device, silicon semiconductor transistors act like tiny light switches that turn a circuit on and off, or amplify weak signals in a communication system. They do this using a small input voltage.

But a fundamental physical limit of silicon semiconductors prevents a transistor from operating below a certain voltage, which hinders its energy efficiency.

To make more efficient electronics, researchers have spent decades working toward magnetic transistors that utilize electron spin to control the flow of electricity. Electron spin is a fundamental property that enables electrons to behave like tiny magnets.

So far, scientists have mostly been limited to using certain magnetic materials. These lack the favorable electronic properties of semiconductors, constraining device performance.

“In this work, we combine magnetism and semiconductor physics to realize useful spintronic devices,” Liu says.

The researchers replace the silicon in the surface layer of a transistor with chromium sulfur bromide, a two-dimensional material that acts as a magnetic semiconductor.

Due to the material’s structure, researchers can switch between two magnetic states very cleanly. This makes it ideal for use in a transistor that smoothly switches between “on” and “off.”

“One of the biggest challenges we faced was finding the right material. We tried many other materials that didn’t work,” Chou says.

They discovered that changing these magnetic states modifies the material’s electronic properties, enabling low-energy operation. And unlike many other 2D materials, chromium sulfur bromide remains stable in air.

To make a transistor, the researchers pattern electrodes onto a silicon substrate, then carefully align and transfer the 2D material on top. They use tape to pick up a tiny piece of material, only a few tens of nanometers thick, and place it onto the substrate.

“A lot of researchers will use solvents or glue to do the transfer, but transistors require a very clean surface. We eliminate all those risks by simplifying this step,” Chou says.

Leveraging magnetism

This lack of contamination enables their device to outperform existing magnetic transistors. Most others can only create a weak magnetic effect, changing the flow of current by a few percent or less. Their new transistor can switch or amplify the electric current by a factor of 10.

They use an external magnetic field to change the magnetic state of the material, switching the transistor using significantly less energy than would usually be required.

The material also allows them to control the magnetic states with electric current. This is important because engineers cannot apply magnetic fields to individual transistors in an electronic device. They need to control each one electrically.

The material’s magnetic properties could also enable transistors with built-in memory, simplifying the design of logic or memory circuits.

A typical memory device has a magnetic cell to store information and a transistor to read it out. Their method can combine both into one magnetic transistor.

“Now, not only are transistors turning on and off, they are also remembering information. And because we can switch the transistor with greater magnitude, the signal is much stronger so we can read out the information faster, and in a much more reliable way,” Liu says.

Building on this demonstration, the researchers plan to further study the use of electrical current to control the device. They are also working to make their method scalable so they can fabricate arrays of transistors.

This research was supported, in part, by the Semiconductor Research Corporation, the U.S. Defense Advanced Research Projects Agency (DARPA), the U.S. National Science Foundation (NSF), the U.S. Department of Energy, the U.S. Army Research Office, and the Czech Ministry of Education, Youth, and Sports. The work was partially carried out at the MIT.nano facilities.

Researchers from MIT have designed a new qubit architecture that enables qubits to interact with each other much more quickly while remaining very stable. This advance could someday help scientists build practical quantum computers that can run long, complex algorithms with high accuracy.

Qubits, which are the building blocks of a quantum computer, usually only store data and rely on other electronics to perform operations and communicate. But qubits are so fragile and error-prone that it is difficult for scientists to connect enough qubits before they lose their information and need to be reset.

The MIT team designed a dual-purpose qubit with two separate parts: one component that stores data and one component that interacts with other qubits and electronics. This design improves the reliability of the qubit and enables it to operate with a reduced error rate, so it can perform more computations in the same time span.

Their simulations indicate that this new qubit architecture could allow significantly faster and higher-fidelity operations than existing designs. 

While this research is still in its early days, it holds the potential to help scientists build large-scale, useful quantum computers that can solve real problems which are too difficult for traditional supercomputers to handle.

“This work feels like a big step. It is a new architecture that shows how much these systems can be engineered. We have taken two ideas and put them together in a way that can help us accomplish this qubit codesign that we are looking for, creating a pretty rare combination of the things we need to do quantum error correction,” says Alec Yen, who earned his electrical engineering and computer science (EECS) PhD this spring and is co-author of a paper describing the new architecture.

He is joined on the paper by lead author Jeremy Kline, an EECS graduate student; Stanley Chen, an MIT undergraduate; and senior author Kevin O’Brien, an associate professor in EECS and principal investigator in the Research Laboratory of Electronics (RLE). The work appears in Physical Review Applied.

A dual-purpose qubit

Just like the bits in a classical computer, quantum bits store information. But unlike classical bits, quantum bits have very short lifespans and can break down quickly when scientists connect them to make a quantum computer.

This degradation, known as decoherence, introduces errors in computations that rapidly build up, derailing long calculations before they are complete.

“The goal for doing all this is to build a fault-tolerant quantum computer where you can correct these errors as they happen, so then you can do long computations and actually do useful things with a quantum computer,” O’Brien explains.

To make qubits more reliable, the MIT researchers developed a new design that includes two separate but connected components: one which stores data and one which interacts with every other part of the quantum circuit.

This interaction component is like an arm that reaches out to other parts of the system, so the researchers call their design the “arm qubit.”

“It is engineered for these two, dual purposes — accomplished together by the data mode and arm mode — and these two goals really matter when you try to do quantum error correction,” Yen says.

Essentially, their design combines two different types of qubits. To make the data mode, they use one popular qubit design which has been known to have a very long lifespan, or coherence. 

The arm mode utilizes a different design that exhibits very strong interactions with other components such as a resonator, which is an electronic component that allows for readout of quantum computations. Readout is the process of measuring a quantum system’s state and translating it into a classical value.

The key to this new architecture is a special coupling unit the researchers previously developed, which they used to connect the data mode and the arm mode. 

Stronger coupling

Normally, coupling the modes together could cause unwanted interactions between them that would build up as more qubits are linked to the system. 

One way to avoid this mixing is to use a technique called nonlinear coupling, which occurs when two components are linked in such a way that changing the state of one causes the other to change in response. Nonlinear coupling is essential for running most quantum algorithms.

The special device the researchers used, known as a quarton coupler, enables very strong nonlinear coupling between the data mode and arm mode, which significantly reduces unwanted mixing. This coupling allows the qubit to perform operations faster before it decoheres.

“By dedicating the ‘arm’ component to coupling, we were able make a design that is scalable, robust to manufacturing errors, and still uses a quarton coupler to achieve strong nonlinear coupling,” Kline says.

When they tested the design in simulations, the arm qubit outperformed other superconducting qubit architectures by yielding state-of-the-art coherence time as well as faster operations and readout. 

The speed and reliability of this new architecture may accelerate quantum error correction, which is an important step in making quantum computers practical.

From here, the researchers plan to work toward fabricating the arm qubit so they can further study its properties and capabilities and integrate it into real quantum systems. 

“This work leaves me with a lot of suspense because our simulations are very promising. Next, we’ll need to see if we can make it, and determine whether we missed anything in the modeling or design. If we can fabricate this qubit, it could be a building block for future error-correcting quantum computers,” O’Brien says.

This work is funded, in part, by the Army Research Office, the Air Force Office of Scientific Research, a Doc Bedard Fellowship from the MIT Center for Quantum Engineering and the Laboratory for Physical Sciences. 

Each year, farmers around the world spend $80 billion on pesticides for their crops. Those pesticides impact not only harmful insects but also bees and beneficial bacteria in the soil. They can also run off into waterways and harm the environment. And, they are increasingly being linked to human diseases like Parkinson’s and cancer.

Amid growing awareness of those problems, pesticides made from living microbes are gaining popularity. Unfortunately, such microbial pesticides are often less effective, forcing farmers to choose between potential environmental damage and higher crop yields.

Now, Robigo is equipping naturally occurring microbes with more potent pest-fighting capabilities. The company, which was co-founded by Andee Wallace PhD ’20, uses technologies more commonly associated with medical applications, like RNA interference and CRISPR, to engineer self-replicating microbes that target plant pathogens more precisely than chemical pesticides and more effectively than other biologically based solutions.

“Chemical pesticides have been a cornerstone of agricultural production for the past 70 years, to the point that it’s nearly impossible to envision an agricultural system without them,” Wallace says. “But that’s the long-term vision we have: providing growers new tools to enable a food system that is in balance with the environment, and that is productive, resilient, and safe.”

In field trials across five states, the company has already shown its microbes offer comparable results to chemical pesticides. In one trial comparing Robigo’s product with another commercial microbial product last summer, Robigo’s system led to a 250 percent increase in crop yield.

“Many crops, like lettuce, are harvested by hand, and the grower told me if a disease reduces yield even by just 25 percent, it’s not economical for them to pay workers to harvest the field at all,” Wallace says. “Growers are just trying to produce enough food to feed everyone. That’s why they use pesticides in the first place. We’re trying to give them a better choice.”

Engineered biology for agriculture

Wallace did her PhD in the lab of Chris Voigt, MIT’s Daniel I.C. Wang Professor and the head of the Department of Biological Engineering. She joined the lab after working at Bolt Threads, a startup spun out of the Voigt lab that was designing a material for the fashion industry inspired by spider silk.

“I came into MIT knowing that I wanted to join Voigt’s lab,” Wallace says. “I was really enamored with biomaterials in general. There are so many examples of animals and organisms that make incredible materials that we humans can’t replicate.”

Wallace’s PhD focused on engineering microbes in an attempt to replicate intricate glass nanostructures produced by single-cell algae called diatoms.

Wallace enjoyed her startup experience and explored entrepreneurship throughout her time at MIT. But it wasn’t until after graduation that she reconnected with two MIT students, Jai Padmakumar PhD ’23 and Connor Sweeney ’21, and decided to start her own company.

The founders’ initial idea was to engineer microbes to deliver CRISPR to target and kill bacteria that are harmful to crops. They used a number of MIT resources to get the company off the ground, including the Venture Mentoring Service, MIT Sandbox, delta v, and the MIT $100K Entrepreneurship Competition. Sweeney was involved in the venture for about a year. Padmakumar left Robigo in 2022.

Today Robigo is addressing a problem of growing importance to the agriculture industry.

“Chemical pesticides are under incredible pressures: increasing scrutiny from consumers and regulators, and increasing pesticide resistance among pests, diseases, and weeds,” Wallace explains. “Over the past 40 years, only two new herbicide chemistry modes of action have been commercialized, so people are understandably worried. If we can’t develop new solutions, resistance is only going to grow and will leave growers without effective tools to protect their crops. I think biotechnology has the potential to solve that problem.”

Farmers hope so, too: In an attempt to address environmental and health concerns, they have increasingly turned to so-called biological pesticide solutions, which are mostly made from natural sources like plant extracts, microbe-derived natural products, and increasingly biotechnology solutions like peptides and RNA.

“They are safer and better for the environment, but currently they just don’t perform as well or as reliably as synthetic chemistry pesticides, so there’s a big distrust among growers,” Wallace says. “Growers are being asked to choose between high performance or safety and sustainability. Robigo is trying to solve that problem by giving them products that do both.”

Robigo’s ARGO biotechnology platform combines synthetic biology and proprietary computational design processes to engineer microbes that perform at a similar level to chemical pesticides, but with improved safety profiles for people and the planet. A key part of that approach is leveraging microbes’ self-replicating abilities to continuously produce and deliver bioactive molecules in the field over the course of the growing season.

The company has moved in recent years from delivering CRISPR to RNA-interference, or RNAi, which inhibits key functions in the pathogens they want to target.

Robigo also differs from other microbial pesticide companies in its approach. Wallace says other companies screen to discover new microbes with the properties they want, then cultivate those for sprays and other modes of applications. But these specialized microbes may not be able to thrive in, say, the microbiome of California farm soil where they’re needed. That means they may die off soon after being deployed. Robigo, conversely, focuses on equipping robust, industry-proven microbes with the ability to target specific pests and diseases.

“Our starting point is ‘What crops will this be used for? and ‘What diseases do we want to control?’” Wallace says. “To design safer products, we need to be direct in how we’re designing RNAi to target different diseases. Another layer of our technology is what we call RNAi stacking, where we combine multiple RNAi into a single microbe to broaden the spectrum of pathogens we can control with a single product.”

Lab to farm to table

Last year, Robigo ran field trials for its two lead products, with soybeans and lettuce across the U.S. Midwest and West. Working with third-party testing companies, they showed a single application of their microbes offered protection for crops over the entire growing season and matched the performance of the leading chemical pesticide at a fraction of the cost. 

“That’s very unusual for biological products, and even many chemical products, so we’re really optimistic about engineered microbes being a new solution that disrupts the conventional chemical pesticide paradigm,” Wallace says.

Wallace says Robigo is expanding fourfold this year and plans to expand even faster next year with the help of major agrochemical companies interested in more sustainable solutions. The company is also partnering to expand to other crops as it helps farmers around the world.

“There are a lot of opportunities we’re excited about, and we’re working with a number of partners as we scale,” Wallace says. “Over the past nine months, we’ve systematically used our ARGO platform to tackle new opportunities, and we have a number of products in the pipeline we’re working to bring to growers around the world.”

How do you build something that lasts? For MIT Assistant Professor Iwnetim "Tim" Abate, the answer is the same whether he’s reimagining how the materials beneath our feet can store energy and manufacture essential chemicals, or mentoring MIT’s future researchers: focus on the foundation.

Rocks provide an unexpected thread connecting Abate’s research and his approach to mentorship. His research brings together electrochemistry, materials science, and Earth sciences to explore how the materials that make up our planet can be harnessed to address some of society’s most pressing challenges in energy and sustainable manufacturing.

In one line of inquiry, his group uses Earth-abundant elements found in rocks, such as manganese and iron, to develop high-energy, low-cost, and more sustainable batteries. In another, they are exploring how the Earth’s subsurface itself could function as a chemical factory. By harnessing naturally reactive rocks, geothermal heat, and injected fluids, they seek to pioneer new ways of producing valuable fuels and chemicals underground, with lower external energy requirements and emissions than conventional industrial processes.

Although batteries and subsurface chemical manufacturing operate at vastly different scales, they share a common philosophy: understanding the intrinsic chemistry of Earth’s materials deeply enough to harness it for useful transformations.

While Abate's research spans a broad range of scientific disciplines, his approach to mentorship is guided by a simple principle: helping students lay the groundwork for their careers after graduate school. Rather than measuring success solely through publications or technical accomplishments, he strives to equip students with the scientific skills, resilience, curiosity, and perspective needed to navigate any path their career may take.

"I often think about mentorship through the image of a rock," Abate explains. "A structure built on rock can withstand storms and the test of time. In the same way, I believe the most important role of a mentor is not simply to help students complete a project or publish papers, but to help them build a strong foundation."

Abate puts this philosophy into practice through his investment in his students' growth as researchers, professionals, and individuals.

In celebration of his exemplary mentorship, Abate has been recognized through MIT's Committed to Caring initiative, a student-driven program that honors graduate mentors who foster supportive and inclusive research environments.

Building holistic relationships

Students often arrive at graduate school with different ambitions. Whether they hope to pursue academia, industry, entrepreneurship, or public service, Abate begins by learning about each person's long-term goals.

Each time a new student joins his group, he meets with them individually to discuss their aspirations and helps tailor aspects of their PhD experience accordingly. Students say these conversations continue throughout their time in the lab, with regular one-on-one meetings focused on both research progress and career development, homing in on their opportunities beyond MIT.

For students interested in entrepreneurship, Abate leverages his own network, introducing them to venture capital firms, philanthropic organizations, and collaborators working across academia and industry. He encourages his students to pursue internships, recognizing that experiences outside the university can strengthen both their research perspective and their future careers.

Students also emphasize his ability to connect them with the expertise they need to push research forward. Whether facilitating access to specialized instrumentation or identifying researchers with complementary knowledge, Abate actively builds the relationships that allow his students and their projects to thrive.

Despite leading a growing research group while balancing teaching responsibilities and launching a startup, nominators wrote that Abate "consistently [shows] up for his students."

He makes time for individual chats with students, subgroup discussions, and weekly lab meetings, all while actively seeking their perspectives on research challenges. "Tim is often curious [to hear] our point of view on research problems and actively looks for our feedback," reflected one nominator. 

This openness creates a synergistic environment where students are encouraged to help shape the direction of the group's work.

Creating space for ambitious ideas

Innovation, Abate believes, depends on more than technical expertise.

"Students need to know that it is OK to pursue ideas that may not work, and that setbacks are part of discovery, rather than signs of failure," he says. "My goal is to create an environment where ambitious ideas are welcomed, careful thinking is valued, and students know they have someone who believes in them through both successes and disappointments."

Students say this philosophy is reflected in the way that Abate approaches advising. Rather than directing every decision, he encourages them to think independently, remaining available whenever guidance is needed. His vast professional network often becomes an extension of that mentorship, opening doors to partnerships and expertise that help students tackle increasingly ambitious research questions.

This commitment to building strong foundations extends beyond his own research group. Since graduate school, Abate has worked to expand access to STEM education through his nonprofit Sci-Fro, which supports educational outreach across Africa. He has also contributed to broader efforts to strengthen scientific infrastructure and research institutions across the continent. 

For Abate, these efforts reflect the same philosophy that guides his mentorship: lasting scientific progress depends not only on individual discoveries, but also on investing in people, communities, and institutions that enable future generations of scientists to thrive.

Supporting the person behind the PhD

Abate regularly checks in during one-on-one meetings, asking how his students are doing and what support they need. He believes these conversations are an essential part of advising.

"Graduate school is one of the most formative periods of a person's life," he says. "While research is important, I don't think success should come at the expense of health, relationships, or personal growth."

He encourages students to build lives that remain meaningful beyond the laboratory, recognizing that the habits, friendships, and perspectives developed during graduate school often shape them just as much as their scientific accomplishments.

Through steady guidance, meaningful connections, and genuine care for each student's well-being, Abate demonstrates a passion for developing exceptional researchers.

"I hope they leave MIT with a strong foundation — both scientifically and personally — that enables them to navigate future challenges, lead with integrity, and build fulfilling lives wherever their careers take them."

The experience of transitioning from research based in theory to focusing on real-world application can vary significantly for different researchers. However, for two former MIT graduate students and a former postdoc, all now at IBM, working with the MIT-IBM Computing Research Lab (formerly the MIT-IBM Watson AI Lab) during their formative years enabled them to not only close the gap between education and employment, but also to generate ideas promising to business impact. 

Despite pursuing varied careers in quantum machine learning, reinforcement learning and artificial intelligence agents,and trustworthy and fair AI, respectively, Srinivasan Arunachalam, Zhang-Wei Hong PhD ’25, and Irene Ko PhD ’24 have consistently found ways to tackle problems defined by novelty and rigor, and translate them to systems with real constraints. Here, the MIT-IBM Computing Research Lab served as a conduit for research relationship building and the flow of their expertise to industry applications.

“Among all the industrial labs, I think MIT-IBM has way better academic collaboration policy and opportunity [than the others],” says Hong, an IBM research staff member with the MIT-IBM Computing Research Lab who began his PhD at MIT in 2020 in the Department of Electrical Engineering and Computer Science (EECS).

Hong has been captivated with reinforcement learning since discovering that DeepMind could play Atari and learn from raw screen pixels via feature engineering. During his graduate work with EECS Associate Professor Pulkit Agrawal, who is also a principal investigator with the lab, Hong sought to build on this: improving value function learning for reinforcement learning in video games, using “Montezuma’s Revenge” in Atari, in order to predict and optimize the policy performance of an agent. With the lab, Hong developed techniques to ground AI for more realistic applications and provide better reward feedback, which he applied to domains such as robotics, large language models (LLMs), and reinforcement learning for science. 

“I’m very excited about curiosity-driven exploration,” says Hong of the MIT-IBM graduate work that helped propel him into his profession. This, he says, allows agents to be inquisitive about new data, like humans, and perform a variety of tasks — from generating test cases to stress-test LLMs to exploring new environments. Now, as a mentor for students of his own, Hong continues to pursue similar lines of open-ended reinforcement learning research, leading him to investigate test-time training for agents and foundation models, and develop infrastructure for IBM’s agentic framework for enterprise tasks like chart reading and tool calling for database queries. This includes evolutionary computing to drive better optimization for exploration and leveraging neuroscience to inform deployment time model improvement. 

“If successful, I think that it would be a very useful system and framework for all of the practitioners in reinforcement learning, because it will be the first framework that enables a model to improve — self-evolve their model weights online at a deployment time,” says Hong.

Irene Ko’s research has also been value-driven, from a personal and professional standpoint. “I started to work [on trustworthy AI] with IBM researchers from day 1 in my PhD, because it was funded by MIT-IBM,” says Ko. This, she says, was particularly advantageous since her goals to develop frontier-safe, robust, accurate, and fair AI also align with that of MIT and IBM, closing the gap between development and real-world deployment. “That really strikes a balance between pure research and something that’s of industry standard or value.” 

Further, her MIT-IBM collaboration through her advisor in EECS, Joseph F. and Nancy P. Keithley Professor Luca Daniel, and IBM Principal Research Scientist Pin-Yu Chen, helped define the direction and parameters of her work to maximize impact, first in neural networks and later with foundation models and LLMs. After graduating in 2024, Ko joined IBM Research to continue her work on trustworthy AI as a research scientist. 

“The reason I chose to go into industry after my PhD, and IBM specifically, is that I found great joy in the collaboration during my PhD. That process, those five years, gave me very high rewards in personal fulfillment,” says Ko. “I wanted to continue the momentum.” 

Her current project focuses on finding pain points in current trustworthy methods that are not widely deployed in AI inference platforms. Unlike using low-rank adapters, which add extra steps to monitor and modify model behavior, her work on vLLM Hook provides a way to access internal model signals, like hidden states or activations, for decoding LLMs. This vector acts on transformer modules to analyze safety scores, such as identifying the likelihood of prompt-injection and hallucination. Here, Ko has developed a lightweight vLLM inference engine plugin framework to program the model internals that could provide significant cost savings over other methods. “I’m very proud of this project because this is really, as far as we know, the first bridge between the deployment and development in trustworthy AI with the inference engines.”

While Srinivasan Arunachalam has always dabbled in quantum research, he constantly explores other areas of theory, seeking to find quantum insights and deep math in unexpected lines of inquiry and papers. “Right off the bat, you don’t see it. You think, maybe this is just a vanilla problem, and then once you start investigating it further, you find some really interesting math that comes out of it, which I think is pretty cool,” he says. 

This drew Arunachalam to MIT as a postdoc in 2018 in the group of Professor Aram Harrow in the Department of Physics. With a learning theory-first perspective, Arunachalam looked for target algorithms, subroutines, and circuits where quantum speed-ups might be possible. Conversations with Isaac Chuang, the Julius A. Stratton Professor in Electrical Engineering and Physics and an MIT-IBM PI, led him to collaborate with the lab and IBM researcher Kristan Temme. 

With a seamless transition to IBM, Arunachalam more closely involved himself with problems that are potentially implementable on a near-term quantum device, keeping in mind constraints like nearest-neighbor architecture, noise, and simpler observable measurements. During this time, Arunachalam focused on quantum machine learning and areas where quantum computing would be superior to classical computing, increasingly prioritizing provability grounded in theory to heuristics. That MIT-IBM connection helped turn theoretical questions into concrete research directions, shaping work that culminated in two prominent papers: one on Hamiltonian learning, which gave rigorous guarantees for learning the dynamics of quantum systems, and another on quantum kernels, which provided theoretical evidence that quantum feature spaces can offer advantages over classical kernels under widely believed hardness assumptions.

Arunachalam also continued to expand his knowledge base by pouring himself into different branches of computer science to uncover structure in problems others may have missed. “One thing which I’ve been a huge fan of is exposing connections between different fields.” This has allowed him to explore learning quantum states — from completely classically simulatable quantum objects to the extremely complicated quantum objects.

Although Hong, Arunachalam, and Ko navigate different domains, they share an instinct: to move ideas across the space between what is possible in principle and what is useful in practice. In their own way, each is applying knowledge gained from collaborations, like that of MIT-IBM Computing Research Lab, to develop “killer applications” — a real-world use case that proves the underlying research can matter beyond the lab.

Self-driving cars are often controlled by deep learning models that sometimes fail in unexpected situations. For instance, the car might inexplicably brake and block the path of an oncoming emergency vehicle. A human driver or passenger may need to react rapidly to prevent a collision.

To help humans better anticipate a vehicle’s mistakes, researchers from MIT and autonomous vehicle technology company Motional developed a new method that provides clear explanations of the underlying model’s decisions.

Usually, the internal reasoning process of a deep learning model is opaque and difficult to understand. But the new method, called the Concept-Wrapper Network (CW-Net), translates that reasoning process into concepts that faithfully describe the autonomous vehicle’s decisions without altering its driving performance.

CW-Net explains the decisions of machine learning-based planners using understandable concepts, like “approaching stopped vehicle” or “close to cyclist.” These explanations can correct misconceptions drivers and passengers have about vehicle behavior and improve their situational awareness.

In road tests on a private track, CW-Net explanations helped safety drivers more accurately predict vehicle behavior; a larger simulation study with nonexpert users yielded similar results. These experiments show how CW-Net can provide important feedback for engineers as they troubleshoot in-vehicle artificial intelligence systems. In the longer term, this technique could boost the safety and transparency of autonomous vehicles, while building appropriate trust in drivers and passengers.

“This work shows how explanations are supportive to the human’s mental model and understanding of the behavior of a system, and how it could be used in engineering and development to improve the technology,” says Julie Shah, an MIT professor of aeronautics and astronautics, director of the Interactive Robotics Group in the Computer Science and Artificial Intelligence Laboratory (CSAIL), and co-senior author of the paper on CW-Net. “Unless we are building these technologies in a way that we can rely on and predict their behavior, then it is a shaky and unsafe foundation for their use.”

She is joined on the paper by lead author Eoin Kenny, a former MIT postdoc who is now a senior AI researcher at J.P. Morgan Chase; co-senior author Momchil Tomov, a staff research scientist at Motional; as well as Motional team members Akshay Dharmavaram, Sang Uk Lee, Tung Phan-Minh, Shreyas Rajesh, Yunqing Hu, and Laura Major, president and CEO of Motional. The research appears today in Nature.

Faithful explanations

Machine-learning-based planners act as the “brain” of a self-driving car. These powerful deep-learning architectures process data from the vehicle’s cameras and lidar sensors, generate a high-level summary of the vehicle’s environment, decide what the car should do next, and output a trajectory for it to follow.

The planners are usually black-box models, which means their internal decision-making process is so complex it is difficult to understand. This can leave scientists and safety drivers in the dark about why an autonomous vehicle made an unexpected decision, like phantom braking.

The researchers designed CW-Net to explain a vehicle’s decisions using understandable concepts, while ensuring those explanations accurately reflect the true reasons behind its behavior. 

“Especially in high-stakes settings like self-driving cars, it’s important that the explanations are not potentially misleading. Because CW-Net is causally faithful in how it makes decisions, that provides certain guarantees around the explanations,” Kenny says.

CW-Net is a “concept classifier,” an AI algorithm that has been trained to predict the high-level concepts that exist within input data. The researchers plug the CW-Net module into the middle of an autonomous vehicle’s existing machine-learning planner architecture.

It translates the model’s internal reasoning process into understandable concepts, like “approaching stopped vehicle” or “close to cyclist.” Then it forces the final piece of the planning model architecture to use those concepts when it decides what the vehicle should do next. In this way, CW-Net ensures the concepts faithfully explain the vehicle’s actions. 

At the same time, CW-Net uses the concepts it classified to generate clear explanations that are output along with the vehicle trajectory, in real-time.

“Instead of just wondering why the car stopped, having real-time data provides feedback that lets you test the system during deployment. You could also give that data to an engineer to potentially improve the system,” Kenny says. 

The researchers trained CW-Net to predict concepts using a dataset of 130 million examples of scenes from self-driving cars, with multiple labeled concepts in each scene. Using such a large, labeled dataset enables it to identify concepts accurately in a wide range of settings.

They also designed CW-Net to mimic the driving decisions of machine-learning-based planners, so the module would not negatively impact vehicle performance.

In the end, CW-Net generates accurate, understandable explanations without altering the original deep learning model.

Improving situational awareness

To test CW-Net, the researchers deployed the module on a real autonomous driving test vehicle (a Motional robotaxi) on a private track with a safety driver. They found that CW-Net helped the safety driver better predict how the vehicle would behave in surprising situations.

For instance, the vehicle consistently stopped when it approached a cyclist, and the safety driver assumed it did so because it detected that cyclist. But CW-Net explanations revealed that the model wasn’t properly configured to detect the cyclist and chose a trajectory that would have caused a collision. Instead, it stopped because its emergency braking procedure kicked in when it got too close.

Armed with this information about the model’s mistake, the safety driver could reduce speed or engage manual driving mode sooner in similar situations. This could also help engineers fix the model to avoid this failure in the future.

In larger online simulation studies using real driving situations captured on the roads of Las Vegas, the researchers saw similar results. CW-Net explanations significantly improved participants’ abilities to predict how an autonomous vehicle will behave.

In the future, the researchers could extend CW-Net so the module can cover more concepts and explore different training and design techniques that could boost performance and improve interpretability.

“Our study shows how crucial interpretability can be to these high-stakes environments, and how it should be on the mind of people as they are making AI in the future, for self-driving cars or other safety-critical environments,” Kenny says.

Neutrinos are the pervasive yet intangible particles that permeate the universe, streaming through whole planets, stars, and our bodies by the trillions each second. The elementary particles are often described as “ghostly” for their near-zero mass and their elusive nature, as they have very little interaction with normal matter. 

Since their discovery in 1956, neutrinos have continued to surprise physicists with their unexpected properties and behaviors. For instance, the particles come in multiple “flavors” and can morph from one to the other like subatomic shape-shifters. Neutrinos may also be their own anti-particle, in a Jekyll-and-Hyde-like quantum duality. And their extremely weak interactions make them nearly impossible to detect.  

Last year, scientists seemed to add to the particle’s mystique, with a concept for a neutrino laser. They proposed that a concentrated beam of neutrinos could be produced by cooling a cloud of radioactive atoms to nanokelvin temperatures, one-billionth the temperature of interstellar space. Slowed to a near-frozen crawl, the atoms would form a Bose-Einstein condensate and should act as one quantum, coherent whole, in a way that speeds up and amplifies their radioactive decay. The physicists assumed that neutrinos, being a natural byproduct of radioactive decay, should also be amplified, and that such a process should emit a laser-like beam of the ghostly particles. 

But work by MIT physicists has now shown that the neutrino laser concept, and a similar proposal for gamma-rays, is impossible. In two companion papers appearing today in Physical Review Letters, Wolfgang Ketterle, the John D. MacArthur Professor of Physics at MIT, together with postdocs Hanzhen Lin and Yu-Kun Lu, presents a two-part analysis that demonstrates both concepts are physically and fundamentally not possible. More specifically, they have shown that the neutrino laser concept is flawed, due to “recoil” (as in, the kinetic energy created by the reaction), and due to a neutrino’s fundamental “fermionic” nature. 

“These two papers are sort of punch one and punch two,” Ketterle says. “Each paper would have killed the proposal.”

MIT professor of physics Joe Formaggio, who put forth the neutrino laser proposal with Ben Jones, who at the time was associate professor of physics at the University of Texas at Arlington, sees the new results as a convincing and constructive challenge. 

“When a new idea — such as the one we proposed — is shared, it is the duty of the community to scrutinize it. Such is the scientific process,” Formaggio says. “Indeed, it was great to see how our paper generated a lot of thinking outside of our original concept. We suspect that will continue.”

A quantum amplifier

The proposal for a neutrino laser was based on the idea of “superradiance” — a quantum, amplifying effect that had only been observed for photons. 

One form of superradiance occurs when a cloud of atoms is cooled to near absolute zero, at which point an atom’s motion is determined not by thermal effects, but purely by quantum uncertainty. In this state of near standstill, which is known as a “Bose-Einstein condensate,” (BEC) the atoms move in sync, as a quantumly correlated whole. 

If photons are pumped into the condensate as a laser beam, the atoms synchronize to scatter the photons back out, in the exact same direction. In contrast, a cloud of atoms at room temperature would simply scatter the photons in random directions, generating, at best, a soft glow. As photons scatter off atoms, the atoms should in turn “recoil,” as if they were physically pushed backward from the impact. In a BEC, because the atoms recoil in sync, the rate at which they scatter photons, in the same direction, grows exponentially. This amplifying effect results in a “superradiant” laser of photons, which scientists have observed. 

In their proposal, Formaggio and Jones, who is now at the University of Manchester, suggested that the same superradiant effect could be possible for radioactive atoms, which naturally emit neutrinos as they decay. If a cloud of radioactive atoms were cooled to form a Bose-Einstein condensate, a similar amplifying effect should kick in and generate a concentrated beam of neutrinos as the atoms decay in sync. To illustrate their point, they outlined a scenario in which a cloud of radioactive rubidium atoms, once cooled into a BEC, would accelerate its radioactive decay, from a half-life of 86 days, to one minute.

No one has ever produced a BEC from radioactive atoms. But if it could be done, then the quantum state should, in theory, produce a neutrino laser. 

Instant recoil

For Ketterle, the idea seemed too good to be true. Ketterle is the leading expert on Bose-Einstein condensates, which he co-discovered in 1995, and for which he shared the Nobel Prize in Physics in 2001. He and his group at MIT have revealed many surprising properties in Bose-Einstein condensates and other ultracold matter, where the energy of atoms is at their lowest.

“My experience has always been that the condensate can do marvelous things at low energy — superfluidity, vortices — and if you were to speak in a room filled with condensate, it would take one hour for you to hear my voice. That’s how slow the condensate is,” Ketterle says. “And I had always come to the conclusion that for anything violent, like nuclear reactions, the condensate would not do anything.”

Compared to visible photons, which have an energy of 1 electron volt, neutrinos are naturally emitted as atoms decay, with a million times more energy. When a neutrino blasts out from an atom, the emission should cause the atom in turn to recoil a million times more strongly than for visible photons. 

“As long as the recoil atom stays in the condensate, it can make the condensate superradiant,” Ketterle says. “But when a neutrino is emitted at a million electronvolts, the atom recoils at velocities equivalent to Mach 10, faster than a fighter jet. This is so fast that the atom would almost instantly disappear.”

Even so, the neutrino laser proposal assumed that the escaped atom should leave a sort of quantum imprint in the condensate, which tells the condensate as a whole to emit future neutrinos in the same exact, laser-like direction.

But in the first of two new papers, Ketterle and his team show through a theoretical analysis that this is not the case. They considered a model that describes superradiance. This model determines the conditions that would lead to superradiance of photons. Ketterle applied the model to the case of radioactive atoms and neutrinos, taking into account the range of energies at which the particles are emitted, as well as the resulting recoil of the decaying atom and the dynamics of the condensate throughout. 

These calculations showed that, in every scenario the team considered, superradiance was not possible. The atom simply recoiled too fast for any quantum imprint to build up. It was as if the condensate instantly loses the “memory” of the neutrino emitted, and therefore would continue emitting neutrinos as atoms normally would, without enhancement.

An anti-memory

In their second paper, the MIT researchers showed that in addition to being impossible due to a physical recoil effect, the concept of a neutrino laser is flawed due to the fundamental nature of neutrinos. 

They found that even if a recoiling atom were to leave a quantum imprint in the condensate, the imprint would not be of what to emit next, but rather, what not to emit. In other words, the memory of the emitted neutrino would tell the condensate to emit the next neutrino in any other direction, preventing the buildup of a directional neutrino beam. The researchers showed that this opposing memory, or “anti-correlation,” is due to the fact that a neutrino is, fundamentally, a fermion. 

Fermions and bosons are the two fundamental classes of particles that make up all the matter in the universe. Bosons are particles with whole-integer spins, such as photons. In contrast, fermions, such as electrons and neutrinos, have half-integer spins. Whether a particle has a whole or half integer spin determines how it interacts at a quantum level with other particles. 

“In superradiance, it is about a memory effect, or quantum correlations in the condensate. And in that context, people had thought that whatever is emitted from the condensate, it doesn’t matter if it is a boson or a fermion,” Ketterle explains. “But we analyzed it, and if you describe it correctly for emitted fermions, you get an anti-memory, which makes the condensate not accelerate in a superradiant form. It rather has the memory to not do it.”

Ketterle, Formaggio, and Jones have met on numerous occasions to talk through the original neutrino laser proposal, and Ketterle’s challenge to it.

“I suspect that someday, someone will do the experiment,” Formaggio says. “Nature, as always, is the final arbiter of such things. And here I would be remiss to not point out that every prior prediction about neutrinos has been wrong. The one thing about neutrinos that never surprises physicists is that they never fail to surprise.”

In part, Ketterle agrees: 

“Creative ideas and discussions among scientists are needed to uncover nature’s surprises,” he says. “But in the case of neutrino lasers, the surprise was too good to be true.”

This research is supported, in part, by the National Science Foundation, the Center for Ultracold Atoms, the Vannevar-Bush Faculty Fellowship, the Gordon and Betty Moore Foundation, and the U.S. Army Research Office.

Walter Torous, senior lecturer in the MIT Department of Urban Studies and Planning (DUSP) and the MIT Sloan School of Management, and director of the Master of Science in Real Estate Development Program (MSRED), was recently named executive director of the MIT Center for Real Estate (CRE) — effective July 1, 2026.

In announcing Torous’ appointment, School of Architecture and Planning Dean Hashim Sarkis also said that Justin Steil, professor of law and urban planning, will represent CRE as faculty chair of the Academic Curriculum Council.

“Together, Walter and Justin will guide CRE’s academic and strategic direction as it continues to strengthen its role within our school and the Institute,” Sarkis said. “Their appointments reflect the center’s distinctive position at the intersection of finance, design, planning, technology, and public policy — and its long-standing commitment to understanding real estate not only as a market force, but also as a driver of urban transformation and social change.”

As executive director, Torous will lead the CRE’s teaching, consortium activities, fundraising, and major events, while continuing to direct the MSRED program. He will also oversee the center’s staff, budget, and strategic direction, and work closely with Steil on the continuing evolution of CRE’s academic programs and industry engagement.

His appointment as executive director follows the tenure of STL Champion Professor Siqi Zheng, who served as CRE faculty director from July 2020 to June 2026.

Before coming to MIT in 2013, Torous was a professor at the Anderson School of Management at the University of California at Los Angeles and founding director of its Ziman Center for Real Estate. In addition to those positions, he also has held faculty appointments at the University of Michigan and the London Business School.

“Since joining MIT, Walter has played an important role in the growth and development of the MSRED program, educating generations of students in real estate finance and mortgage securitization,” Sarkis says. 

“Real estate, both commercial and residential, is undergoing a tremendous change in the U.S., as well as in Europe and Asia,” Torous says. “Demographic changes, as an aging population stays longer in their homes, are creating an imbalance in residential real estate markets. New technologies and work from home are buffeting commercial real estate. Retail is changing.  We’re in a period of turmoil, and I view the center’s role as being primarily to educate the next generation of leaders, especially in technology and financial markets, which are becoming ever more important to the functioning of real estate assets and markets. That requires that we train our students to be very facile with technology, so that they’re not affected by the ebbs and flows of changes, can maintain a strong career trajectory, and be stewards of the real estate industry going forward.” 

For this reason, he would like to see the MSRED curriculum expand beyond DUSP to add more content from architecture, civil and environmental engineering, the Media Lab, MIT Sloan, and other areas of the Institute. 

Torous also wants to more fully engage the 1,200-plus alumni from the center’s 43 years educating graduate students.

“A lot of our alums have assumed important positions in the real estate industry around the world,” he says. “In terms of training the next generation of real estate leaders, there’s a lot that we can learn from the industry leaders we’ve already produced.”

An economist and expert in the financial aspects of real estate known for his empirical studies of derivatives, options, mortgages, and other debt instruments, Torous’ research interests include the reorganization of financially distressed firms and statistical issues in finance.

His recent research has focused on better understanding why homeowners default on their mortgages. He is also interested in the application of machine learning to investigate how the dynamics of the U.S. commercial office market changed with the Covid-19 pandemic, and the lessons developers can learn about the new office market landscape. This research reflects the growing importance of AI and large language models to every aspect of real estate decision-making. Because of this, the MSRED curriculum now includes a class on AI and real estate, and Torous and Steil plan to add other, similar offerings.

“It’s important going forward that we focus on all aspects of real estate,” he says. “I look forward to working with Justin to create a curriculum that goes across the Institute and that will prepare CRE students to be leaders in the field.”

A new theory, published in The Journal of Neuroscience by three scientists in The Picower Institute for Learning and Memory at MIT, offers an explanation of how the brain produces cognition and consciousness: It uses traveling waves of rhythmic neural activity to coordinate nimble neural networks with analog computations. 

The metaphor that the brain operates with “circuits” is incomplete, says Picower Professor Earl K. Miller, the paper’s senior author. Indubitably, the brain’s physically connected circuits provide the infrastructure to store our memories and represent our ongoing needs and goals. But when we need to make improvised use of that knowledge in the rapid-fire, anything-goes sensory context the world constantly throws our way, we can’t just depend on the relatively slow chemical process of rewiring those circuit connections called “synapses,” he says. 

Instead, the brain needs a control system that can coordinate millions of neurons to process information in a fraction of a second. Brain waves, long understood to be the synchronized rhythmic fluctuations of large groups of neurons, turn out to be performing that crucial service, Miller and his colleagues argue, citing years of experimental evidence from his lab and many others.

“Circuits and synapses are important and fundamental, that’s the start. But there is more going on,” says Miller, a member of MIT’s Department of Brain and Cognitive Sciences faculty. “The brain generates waves, and wave dynamics are a highly efficient way to coordinate and perform computation.”

While digital circuits make calculations one step at a time through sequential switches and gates, analog computation, which can be performed via the interference of waves, processes multiple calculations in parallel. That’s not only more efficient, but also locally focused traveling waves happen to be a ubiquitous feature of the brain, the scientists note.

“The brain exploits its own physics,” wrote Miller and co-authors Scott L. Brincat and Jefferson E. Roy, who are research scientists in Miller’s lab.

The new theory is important not only because it provides an explanation of cognition and consciousness, but also because it asserts the potential importance of considering waves in clinical treatment. Conveniently, waves can be manipulated non-invasively.

“Developing treatments based on brain wave dynamics is not just an opportunity, but also an obligation,” says Miller, whose lab is part of a collaboration studying brain waves in autism.

Building the analog argument

To make the case that the brain uses waves to coordinate neurons to produce cognition and consciousness, the scientists begin with the now well-established observation that many neurons don’t just do one job. Instead, they respond to multiple cues and contexts, essentially participating in multiple functional networks at once, a property called “mixed selectivity.” Miller and colleagues have argued for years that this gives the brain immense computational horsepower, but it also initially raised the question of how the brain organizes these multiple overlapping networks with such speed and flexibility to produce the nimble thought we all depend on.

After numerous studies, the answer that has emerged for Miller and many other neuroscientists is that brain waves organize neural ensembles to process information. Miller has shown that brain waves of different frequencies govern cognitive processes such as working memory and predictive coding. Relatively slow “alpha” and “beta” frequency waves, representing memories and goals, regulate faster frequency “gamma” waves, which represent and report incoming sensory information.

The new theory posits that these alpha/beta control waves emerge from the coordinated spiking of neurons in circuits (connected at junctions called “synapses”) that encode stored memories and goals. 

“Synapses store representations, while wave dynamics help determine which representations are active at any given time,” the authors wrote.

In some of the Miller lab’s newer research, the team has found evidence that even as waves emerge from neural spiking, the waves can rapidly grow to directly influence and coordinate spiking via an electric field-mediated process called ephaptic coupling. Importantly, electric fields can exert this coordinating influence very rapidly.

Another essential component of the theory, which Miller’s lab has also shown experimentally, is that alpha/beta waves are capable of exerting their control spatially, by affecting local areas of the cortex, and temporally, by traveling along the cortex. Essentially, the beta waves act as mobile stencils that govern where and when gamma waves can process sensory information and which ensembles of neurons will participate. Taken together, this suggests that the brain engages in “spatiotemporal computing,” the authors write. And where the waves intersect, they can add and subtract, enabling analog computations.

Miller acknowledges that his lab’s next step should be to provide direct evidence that the analog computations are taking place.

“This is a theory. We aim to test it by looking for signatures of analog computation in brain wave patterns,” Miller says. 

Connection to consciousness

The article asserts that consciousness “emerges when these dynamic wave patterns bring the cortex in an organized, globally integrated state, one that naturally links and influences widespread activity.”

Some of the most compelling evidence linking wave dynamics to consciousness comes from studies of general anesthesia that Miller has conducted with Picower Institute colleague Emery N. Brown, who is an Institute professor at MIT, an anesthesiologist at Massachusetts General Hospital, and a professor in Harvard Medical School. Their labs have shown that three different drugs, each with different molecular mechanisms of action, all similarly disrupt brain wave dynamics to produce unconsciousness.

“Consciousness depends less on specific receptors or cell types and more on the integrity of large-scale wave organization,” the authors write in the review.

In other words, much like cognition, consciousness depends on how the brain efficiently organizes itself with brain waves.

“Electric field dynamics offer a low-overhead substrate for organizing and coordinating information across cortical networks,” they conclude. “Given strong evolutionary pressure to maximize computation per unit energy, it would be surprising if evolution did not exploit such a built-in analog computing substrate.”

The Freedom Together Foundation, The Picower Institute for Learning and Memory, the U.S. Army Research Office, the U.S. Office of Naval Research, a MURI grant, the National Institutes of Health, and the Simons Center for the Social Brain supported the research.

A region of the brain called the striatum is critical for many cognitive and motor functions, including decision-making, control of movement, habit formation, and processing of reward. It also plays a role in addiction and is significantly affected by Huntington’s disease, schizophrenia, and other disorders.

In work that could help scientists devise new treatments for those diseases, MIT researchers have generated a new atlas of the neurons found within the striatum. Using single-cell RNA sequencing and other techniques, they were able to identify 31 subgroups of neurons based on which genes they express.

These groups include neurons that are involved in addiction, depression, and schizophrenia. The researchers also discovered why some neurons of the striatum are more vulnerable to Huntington’s disease. All of these results, the researchers say, could help scientists develop new drugs to combat these conditions.

“We see this as the foundation that will allow more studies in our Huntington’s disease and opioid use disorder projects. We needed a roadmap of what is there,” says Myriam Heiman, the Picower Professor of Neuroscience and director of MIT’s Picower Institute for Learning and Memory.

Heiman; Manolis Kellis, a professor of computer science in MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and a member of the Broad Institute of MIT and Harvard; and Dana Gabuzda, a principal investigator at Dana-Farber Cancer Institute and a professor of neurology at Brigham and Women’s Hospital and Harvard Medical School, are the senior authors of the study, which appears today in Cell. MIT postdoc Raleigh Linville and MIT graduate student Benjamin James are the paper’s lead authors.

Mapping the striatum

The striatum, located deep within the brain, receives diverse inputs from the cortex, midbrain, hippocampus, and other regions, which it uses to coordinate planning, movement, and decision-making, as well as processing reward. In this study, the researchers focused on the most populous cell type in the striatum, a type of inhibitory neuron called the medium spiny neuron, which responds to dopamine.

Most of these medium spiny neurons belong to either the direct pathway, which helps to promote movement, or the indirect pathway, which suppresses unwanted movements. These pathways are distinguishable by what type of dopamine receptor they express — dopamine receptor 1 (D1) or dopamine receptor 2 (D2).

Beyond these two divisions, scientists knew that there were many subpopulations performing different roles, especially in the anatomically ventral (lower) regions of the striatum. However, it has been difficult to generate a consensus on how to classify these cells, in part because prior studies focused on specific subregions, meaning that overarching principles of striatal cellular organization were lacking.

To overcome that challenge, the researchers worked closely with brain banks in the United States and Canada to collect postmortem striatal samples representing diverse anatomical regions. 

Then, they used three different techniques to analyze the samples, including single-cell RNA sequencing — a method that can measure RNA molecules within individual cells to reveal which genes are being expressed. Two additional techniques — multiplexed fluorescent in situ hybridization and spatial transcriptomics — allowed the researchers to identify spatial principles of organization within the tissue.

Using these techniques, the researchers were able to identify 31 different subpopulations of neurons, including nine types of medium spiny neurons. Among their medium spiny neuron types are two “outlier” populations that appear to play important roles in schizophrenia, substance use disorder, and depression.

One of those populations, known as D1 outliers, showed high expression of genes involved in addiction and substance use disorder, especially genes related to opioid response. Another population, called D2 outliers, showed high expression of genes that respond to antidepressants. And, both populations appeared to respond strongly to clozapine, an antipsychotic drug used to treat schizophrenia.

Clozapine is among the most effective antipsychotics available, but it’s not widely used in the United States because it can cause a fatal blood disorder in a small percentage of patients. Now that researchers know which cells the drug acts on, they may be able to design more targeted therapeutics to overcome psychosis, but without the harmful side effects, Heiman says.

Huntington’s vulnerability

Another key finding of the paper helps to shed light on why the dorsal (upper) part of the striatum is more vulnerable to Huntington’s disease. The disease is caused by an inherited version of the huntingtin gene that carries too many repetitive DNA segments, called CAG repeats. 

The researchers found that dorsal populations of medium spiny neurons express higher levels of the genes MSH2 and MSH3, which play a role in increasing the number of CAG repeats found in the huntingtin gene. As more of those repeats accumulate, the mutated version of the huntingtin protein becomes more harmful to cells.

The researchers also found that a rare population of medium spiny neurons that forms island-like structures in the ventral striatum was more resistant to the accumulation of CAG repeats. Further study of this class of cells might help researchers learn how to induce other medium spiny neurons to become more resistant to the disease, Heiman says.

“Looking at the genes that these neurons express or don’t express might give us some clues as to how to make other medium spiny neurons resilient like them,” she says. 

Insights into substance use disorders

The researchers also compared their findings from human tissue samples to samples from mice and found several differences, especially in the expression of genes related to drug response and substance use disorders. One such gene, which encodes the mu opioid receptor (OPRM1), is highly expressed in the human D1 outlier population, but not in the corresponding population of neurons in mice. 

This means that standard mouse models may not fully capture the biology of opioid responses, and that engineering mice to express this receptor in a similar manner to humans could make those models significantly more accurate.

“Some of the diversity we’re seeing in the human ventral striatum is species-specific and has implications for modeling substance use disorder in rodents,” Heiman says. “Now that we understand better the species differences, we can use the rodent models for specific questions that apply for conserved genes, but we could also think about humanizing some models.”

The researchers hope that this map, built from tissue contributions by brain donors and their families, and assembled across disciplines and institutions, will provide an important starting point for researchers pursuing new treatments for some of the most difficult-to-treat brain disorders.

The research was funded, in part, by the National Institutes of Health, the G. Harold and Leila Y. Mathers Charitable Foundation, the Freedom Together Foundation, the Natalia Mental Health Foundation, the Biswas Family Foundation, and the Milken Institute.

Before Ila Kumar thinks about how to build technology, she asks a different question: What is the context that technology will operate in, and who needs to be involved in the design? 

For Kumar, meaningful innovation doesn’t result from engineers or designers working in isolation. Instead, she believes the best innovations emerge when the people who stand to benefit from a technology help create it from the very beginning. 

That philosophy has guided her research in the Lifelong Kindergarten group, where she works alongside young people who have experienced trauma during childhood, particularly those involved in the child welfare system, to reimagine how technology can support healing, connection, and independence. 

“I really think that community-based design is the only way that we can make technology that accounts for communities’ needs, but also their barriers, their cultures, their concerns,” Kumar says. “It’s the only way that we can make really sustainable and positively impactful technology.” 

Today, Kumar is preparing to enter the sixth and final year of her PhD. But when she first arrived at MIT in 2021, she envisioned staying only long enough to complete a master’s degree. However, Kumar quickly fell in love with her work and decided to stay at MIT and pursue her doctorate.

Before graduate school, Kumar grew up in Philadelphia, attended the University of Pennsylvania, and worked on several projects at the intersection of technology and mental health or psychology research. 

Through those experiences, Kumar began to question whether the technology she was helping to develop was having the sustained impact she hoped for. “I had done a number of projects that were ‘tech for good,’” she says. “And I wasn’t seeing that what I was doing had a long-term impact.” 

Rather than walking away from technology altogether, Kumar began to rethink how it was created. “If we design technology in community-based ways and really think about holistic well-being,” she says, “maybe we can actually create things that help people.” 

That conviction eventually became the foundation of her doctoral research, and over the course of her PhD, Kumar has increasingly moved from simply listening to communities to building with them. 

Public conversations about technology often present a choice: Embrace it or reject it. Kumar believes that it’s not that simple. 

Kumar sees the way digital platforms have the potential to both harm young people’s mental health and development, and help young people process emotions, strengthen relationships, and practice healthy vulnerability — if those tools are designed thoughtfully and embedded in the systems where young people already receive care and support. 

Much of her work explores exactly what that could look like. 

One project Kumar worked on, in partnership with Stepping Forward LA and with the input of the young people who would use the app, replaces text-heavy communication with a visual collage system to help young people impacted by trauma and the child welfare system to express emotions that may be difficult to put into words, and to build a sense of connectedness with one another. 

In an ongoing project, Kumar is collaborating with the Justice Resource Institute to design a mobile app that supports youth in playing an active role in their treatment-planning process and helps them work toward the goals they set outside of the therapy office. The group is working with clinicians and youth to design and evaluate the system.

“We are not sitting at MIT designing tools and just throwing them at people,” Kumar says.  “We’re designing it together. We need to actually have the folks that are relevant to providing the care in the room.” 

That idea became even clearer to Kumar through a 10-month technology leadership circle she co-facilitated with Foster America. The program brought together people with lived experience of foster care and technology experts to envision how digital technologies could fill gaps in care for young people in the child welfare system. 

This project surfaced the importance of not just designing tools that center youths’ needs but also considering the ways in which social services need to be brought into the innovation process. 

Those ideas have also led Kumar to explorations that involve one of technology’s newest frontiers: artificial intelligence. She began asking questions after she realized that young people who had experienced trauma had already been turning to AI to make critical life decisions, even as many caregivers were not aware of it.

As a result, Kumar has increasingly focused on supporting care providers in talking with young people about AI. She has led training workshops with organizations that serve young people impacted by trauma or involved in the child welfare system.

Kumar’s passion for advocating for young people extends far beyond the lab. She also volunteers as a court-appointed special advocate, working one-on-one with a young person in the child welfare system while pursuing her PhD. 

The role has deepened both her understanding of the challenges young people face and her belief that lasting change depends on relationships.

Some of her most meaningful moments have come while working directly with young people.  Last summer, she, alongside another graduate student in her lab, mentored two interns with foster care experience during a six-week program that blended technology, creativity, and personal growth. 

“It felt like a real privilege,” Kumar says. “Even the six weeks was not enough.” 

Those relationships have also inspired Kumar to address how community-based research is conducted at MIT. 

Recognizing that many students interested in community-engaged work often feel isolated, she collaborated with the Priscilla King Gray Public Service Center to co-teach a course on community-driven innovation. She later established a biweekly community of practice connecting researchers across MIT and Harvard University who are navigating the benefits and challenges of conducting research alongside communities rather than simply studying them. 

Outside of research, Kumar enjoys birdwatching, cooking with friends, and creating graphic illustrations — creative pursuits that, much like her research, reward patience, observation, and careful attention. 

As technology becomes increasingly woven into young people’s lives, Kumar hopes innovation will move beyond the lab and into the communities it is meant to serve. 

“The future of actually impactful technologies,” Kumar says, “is when researchers are making decisions with communities instead of for them.”

Solving complex social problems with multiple interrelated causes can involve juggling a variety of factors. Securing funding, designing the right programs, and sustaining the political will necessary to implement them demands a targeted approach.

Lyonel Tanganco, a graduate student in MIT’s Master in Data, Economics, and Design of Policy (DEDP) program, seeks to connect data, policy, and practice-based community interventions to improve living conditions and service delivery in middle-income countries. His studies have allowed him to work with innovative practitioners making real improvements in the world, he says.

“There’s innovation at work in middle-income countries,” says Tanganco, a native of the Philippines. “Seeing the attitudes to adopt and scale new policies and procedures to improve lives has been very interesting to me.” 

Taking those innovative practices and investigating their adaptability and potential to scale is at the heart of Tanganco’s research and work. “How do we make growth broad-based and inclusive?” he asks.

The DEDP master’s program, jointly run by MIT’s Department of Economics and the Abdul Latif Jameel Poverty Action Lab (J-PAL), equips development professionals from across the globe with the practical skills and theoretical knowledge needed to tackle these and other kinds of challenges. J-PAL seeks to reduce poverty by ensuring that policy is informed by scientific evidence — conducting randomized impact evaluations; helping governments, nongovernmental organizations, donors, and the private sector apply the resulting evidence to their work; and training researchers, policymakers, practitioners, and donors to generate and use that evidence.

Designing a path to more effective policies and practices

Before arriving at MIT, Tanganco earned degrees in management science and economics, graduating at the top of his class from Ateneo de Manila University in the Philippines. He was previously the director of the Policy, Research, and Liaison Office in the Philippine Department of Finance. His work focused on helping develop the nation’s response to the Covid-19 outbreak, tax policy reform, and communications support for key policy initiatives.

“During my time in government, we sought to increase revenues for health care and increase outlays for health-care programs,” he says. “We were thinking about health care from the financing perspective.” 

Tanganco’s efforts helped increase taxes on cigarettes, vaping, and alcohol products, which funded a sixfold increase in the health-care budget. Allocating more funding for health care, he says, may yield better outcomes. 

Additionally, Tanganco supported reforms to increase taxation on top Filipino income earners while lowering taxes for others, which the government subsequently implemented. Later, he and some of his colleagues formed a “policy think-and-do tank” — Malusog at Matalinong Bata Coalition (Smart and Healthy Kids Coalition) — that collaborates closely with government agencies on large-scale social programs. 

There, he played a key role in designing and advancing a conditional cash transfer program aimed at addressing malnutrition that now reaches more than 190,000 Filipino households. “The program gives families the equivalent of $12 per month under the condition that they bring their children for regular monthly checkups,” Tanganco says. “It increased health-seeking behavior eightfold.”

While he saw success in implementing these programs, Tanganco still found gaps in both knowledge and implementation he thought he could close by enrolling in a program like DEDP. “I wanted a graduate program that taught me what I couldn’t get from a professional career,” he says.

Expanding research into targeted areas

Tanganco describes living in a middle-income country as “living in two contradictory worlds at the same time.” 

“I’ve seen gleaming metropolitan skylines alongside underserved communities; pockets of affluence surrounded by persistent poverty; world-class hospitals alongside children who still lack access to basic health care,” he says. “The through line in my work is figuring out how to help middle-income countries translate economic growth to better lives and better human outcomes.” 

His DEDP studies have taken him to Indonesia this summer for work on a capstone project with economist Benjamin Olken, the TEPCO Professor of Economics and co-faculty director of J-PAL. The research, conducted in collaboration with Indonesian local governments, involves the design and rollout of a randomized evaluation of a tax intervention. 

“So far, I’ve visited and conferred with several local Indonesian governments to assess tax administration issues,” he says. Investigating Indonesian governmental interventions may help improve service delivery and support. One of the ways Tanganco hopes to help Indonesians, Filipinos, and others is by developing tools to raise revenues in simple, effective, and fair ways, making it easier to improve constituent sentiment and service delivery. 

Tanganco wants to help policymakers and others understand how politics and other factors influence areas like investments in nutrition and environment. His studies have sharpened his investigative approach in these critical areas.

In the Philippines, for example, one-in-four children is malnourished. “Children who lack proper nutrition before age 2 develop smaller brains, perform worse in school and work, and are far more likely to remain in poverty,” Tanganco reports. “Their potential is capped before they get the chance to use it.”  

Middle-income countries also suffer disproportionately from climate-change-related impacts. “Typhoons and extreme heat severely disrupt learning and economic growth in the Philippines,” Tanganco says. “More than a tenth of school days are lost because of climate issues.” 

Essentially, without improved policies and practices alongside a sustained effort to improve lives, “we’re losing extraordinary opportunities for human advancement to wasted potential,” Tanganco believes. “Experiences like that abound,” he says. 

From the classroom to the next chapter 

Tanganco values opportunities to range beyond his DEDP studies. He fondly remembers completing a doctoral-level course in environmental economics co-taught by Olken and Jacob Moscona, the 3M Career Development Assistant Professor of Economics. Its focus on research appealed to him. “I was glad to have time to think about the problems I’m trying to solve,” he says.

Tanganco also enjoyed exploring Greater Boston with his wife — a graduate student at Harvard University — and his fellow DEDP students. From restaurants to concerts with other music nerds, he appreciates the time they spent outside the classroom. “We discuss our hopes and our home countries’ challenges,” he enthuses. “I’m excited to see what folks will do after this.”

Tanganco is especially pleased with the Institute’s commitment to ensuring scholarship centers an interdisciplinary approach. He likens the MIT educational style to “Avatar: The Last Airbender’s” Uncle Iroh, who recommends drawing wisdom from a variety of elements to ensure wisdom doesn’t grow stale. 

These and additional opportunities to step outside his previously defined areas of expertise left a lasting impact on him. “Everyone at MIT is open to collaboration,” he says. “There are a lot of thinkers and doers here, and you don’t have to work hard to convince other students to help you.”

As Tanganco continues his work, he encourages practitioners — doctors, nutritionists, and community health workers, for example — to partner with economists and other researchers to translate their expertise into quantifiable metrics policymakers can understand. “Develop an eye for impact,” he adds.

Enrolling in the DEDP program “has been game-changing,” Tanganco concludes. “The program provides a solid foundation for understanding the world and how to make a positive, measurable difference in the lives of other people, especially the least fortunate among us.”

After extensive modifications over the past seven years, the Gulfstream IV (G-IV) aircraft operated and maintained by MIT Lincoln Laboratory's Tactical Defense Systems Group and Flight Test Facility (FTF) recently flew home from Canada. 

Transforming the standard business jet into a highly specialized research platform — which will support the U.S. Air Force's Air Vehicle Survivability Evaluation (AVSE) program for decades to come — represented the largest and most complex airborne test bed modernization in Lincoln Laboratory history. The Tactical Defense Systems Group, assisted by the FTF, coordinated the effort with the Toronto-based aerospace company Field Aviation.

"Our team made hundreds of trips to Canada and dedicated countless weekends to keep the project moving along," says David Culbertson, FTF manager. "Seeing the aircraft finally return to the laboratory invoked a sense of pride and satisfaction."

An airborne testing infrastructure

For more than 40 years, the Tactical Defense Systems Group has supported the AVSE program, leveraging airborne test beds to assess how U.S. aircraft and space assets fare against current and emerging threats. The group had been conducting airborne testing for the AVSE program with a modified Gulfstream II (G-II) since the early 1990s. In 2013, they began a series of studies to replace the G-II because parts availability issues were looming. These studies concluded that the G-IV was the best option, given its performance and capabilities, including its respectively higher altitude and longer range; long-term sustainability; and cost. The laboratory purchased the G-IV in 2015.

To avoid repeatedly reopening the costly Federal Aviation Administration (FAA) certification process over the planned operational lifetime of the G-IV (25 to 30 years), the group decided to complete all anticipated aircraft modifications at once, rather than in phases. Following a competitive bidding process, the laboratory selected Field Aviation to perform the modifications. Field Aviation had modified the G-II, in addition to other laboratory aircraft. In December 2018, FTF pilots flew the G-IV to Toronto, where it was expected to remain for approximately three to four years.

However, Covid-19 pandemic-related disruptions and contractor management shifts extended this timeline. To help bring the aircraft home, the laboratory stepped in to oversee aircraft modifications, maintenance, and reassembly. Laboratory engineers, mechanics, pilots, program managers, and legal teams worked together to secure Canadian work permits and maintain a continuous onsite presence. Senior aircraft mechanic Craig Rowe served as lead crew chief, traveling monthly with team members to Canada; for his efforts, he was recognized with a 2026 MIT Excellence Award for Outstanding Contributor. 

A structural overhaul

To modify the aircraft, mechanics removed, tracked, and ultimately reinstalled more than 2,000 components. The revamped G-IV incorporated 12 major modifications that required sweeping structural changes.

For example, on the wings, mechanics installed four pylons for carrying external sensor pods weighing anywhere from 200 to more than 1,000 pounds. The wings had to be structurally fortified to withstand the added weight, stress, and aerodynamic loads that would be experienced during flight. They added a fifth sensor pylon, capable of holding up to 2,000 pounds and accommodating systems nearly 19 feet long, to the forward lower fuselage. Development of the pylons spanned nearly five years because of intensive reverse engineering, including purchasing and disassembling a wing from a scrapped G-IV to measure the internal structural components. Installation took almost two years because access to the inner wing structure was limited to small panels normally used for inspections.

Mechanics modified the roof and lower fuselage to create flat surfaces to allow rapid mounting of external antennas and sensor systems without repeated incursions into the aircraft's pressurized fuselage. They extended the aircraft's nose and tail with standardized sensor-mounting interfaces to enable rapid placement of sensors for both forward- and aft-facing test scenarios. The six-foot nose extension required completely gutting the cockpit so the internal structure could be reinforced to bear the weight of the mounting interface and test systems.

In the interior, the team installed 14 equipment racks; workstations for six onboard operators; fiber-optic, Ethernet, and coaxial cables; liquid- and air-cooling systems; and dedicated power-distribution infrastructure separated from the baseline aircraft for safety reasons.

The remodel also required developing a means to generate sufficient electrical power to operate the test systems in flight while meeting FAA fire-containment standards. The aircraft’s original auxiliary power unit (APU) — normally intended to assist only with engine startup — was far too small for the mission requirements and could not operate airborne. Field Aviation engineers designed an entirely new fireproof titanium enclosure to house a larger APU capable of producing nearly double the original electrical output up to the 45,000-foot G-IV altitude ceiling. The laboratory's Engineering Division ran simulations to validate that the APU inlet airflow would allow for maximum APU power output throughout the flight duration.

Steps toward mission qualification 

After reassembling the G-IV, FTF mechanics conducted hundreds of operational checks to ensure every aircraft system disturbed during the modification worked properly and to validate aircraft safety and readiness to resume flight operations. The aircraft completed multiple post-modification flights without a single maintenance write-up.

"It's extremely rare for a heavily modified aircraft of this complexity to have no write-ups," says program manager Paul Mancini from the Tactical Defense Systems Group. "That's a testament to the quality of work of the FTF mechanics who put the airplane back together and the Field Aviation engineers who completed the modifications."

Since the G-IV returned home this spring, test pilots have been evaluating its airworthiness — i.e., in-flight safety and functionality. The Tactical Defense Systems Group expects approximately another 18 months to complete flight testing, mission systems modification, test systems installation, and FAA certification before the aircraft becomes fully mission-qualified to operationally support the AVSE program.

MIT President Sally Kornbluth formally welcomed the undergraduate Class of 2030 to campus on Sunday, noting that the Institute quickly “feels like home” to new students. 

The annual event, officially called the President’s Convocation for First-Years and Families, is held at the Johnson Ice Rink on campus on the weekend most new undergraduates arrive on campus. 

The Class of 2030 consists of more than 1,100 first-year undergraduates from all over the map, representing a broad variety of academic interests and backgrounds. Yet even for such a wide-ranging group, Kornbluth observed, “It is very, very common for new students to say that in coming to MIT, they have finally found their place. They have finally found their people. And it feels like home.”

Kornbluth’s remarks outlined some of the binding forces that connect students, through the shared culture of inquiry and discovery at MIT.

“I was struck right away by the wall-to-wall enthusiasm for fundamental science, what we like to think of as curiosity on a mission,” Kornbluth said. “Every day here, hundreds of people are pushing the boundaries of human knowledge.” 

This month alone, she noted, “astronomers here just discovered an entirely new type of astrophysical object, a black hole star. … And then, two days later, an MIT research team discovered that a drug that blocks a certain enzyme can reduce the risk of developing lung cancer.” 

Kornbluth added: “And that’s just a regular [occurrence] here. As you’ll see, the discoveries just keep on coming in everything, from climate science to computer science, nuclear science to neuroscience, from chemistry to quantum.” 

Secondly, Kornbluth said, people in the MIT community are frequently motivated by a desire to have an impact through their work.

“We’re also driven to make a positive difference in the world,” she told the audience of more than 2,000, which frequently applauded at key junctures. 

A third common feature of campus life, Kornbluth told the crowd, is the “spirit of entrepreneurship” on campus, generally defined as a propensity to take action. 

“Now, I don’t mean that everybody has to start a company, though a lot of people do,” Kornbluth said. “But at MIT, when we talk about entrepreneurship, we also mean the broad spirit of, do something, try something, with your whole heart … and let the doing teach you how to make a difference.” 

Kornbluth also made a series of remarks about AI, noting that MIT has “deep ties” to the development of the technology and that AI tools are expanding and accelerating work in many fields of research. 

That said, she added, “As educators, it is our challenge to derive AI’s benefits and counteract its harms.” And she called a recent report MIT has issued about AI and education “a powerful reminder that MIT was founded to help human beings develop their own powers of discovery, problem-solving, and invention. That is still and will always be our essential work. It is the experience you all came here for.”

All told, Kornbluth said, “We’re so glad and so grateful that you chose to bring your talent, your energy, your curiosity, and your creativity to MIT. And we’re thrilled to be starting this new year with all of you.” 

Kornbluth then introduced the audience to other campus administration leaders who were sitting onstage for her remarks: Provost Anantha Chandrakasan, Chancellor Melissa Nobles, and Vice Chancellor for Graduate and Undergraduate Education David L. Darmofal. 

Attendees also heard remarks from two faculty members who are also alumni, per convocation tradition. 

Anna Huang SM ’08, the Robert N. Noyce Career Development Professor in both the Music and Theater Arts program and the Department of Electrical Engineering and Computer Science, discussed her work as well as the student experience on campus. 

Huang studies human-computer interactions and develops human-AI collaborations in music making, and urged the students to follow their interests — which, in Huang’s case, are quite broad. She spent years working at Google and is also a composer herself.

“You’re going to discover so much here at MIT,” Huang said. “I discover something new every day.” 

She urged students to participate in campus activities and to pursue programs such as MISTI, the global experiences program at MIT that enables internships, study abroad, and more. Huang also emphasized that MIT is a collaborative, interdisciplinary place where students can thrive by working with others. 

“MIT is a very, very supportive environment,” Huang added. “And we value the perspective and the combinations of unique interests you bring.” 

Huang was followed at the podium by Desirée Plata PhD ’09, associate dean of engineering, School of Engineering Distinguished Climate and Energy Professor, and associate professor of civil and environmental engineering, who urged the students to cultivate an ethos of optimism about their studies and ability to improve the world. 

Plata’s wide-ranging work applies chemical engineering to climate issues — for instance, as she noted, by working to replicate methane-capture processes observed in nature onto new technologies that could be located in mines. Deploying such techniques to reduce the presence of greenhouse gases could help slow the worldwide rise of temperatures. 

“Modulating the warming rate of the planet is admittedly ambitious,” Plata said. “But it’s not impossible. At least not from a thermodynamic perspective. And that’s just the kind of problem we like to solve.” 

Plata also encouraged students to cultivate a feeling of open-minded optimism about their own pursuits.

“When I walk onto MIT’s campus each morning, I take a deep breath. I feel that same sense of possibility that I felt the [first] time I set foot here,” Plata said. “A high privilege of my life is being able to engage some of the most talented minds of our time. To engage all of you. To help develop your respective paths. And enjoy the amplifying impact you’re going to go on and have in this world.”

After Plata spoke, Kornbluth, who is from a musical family and enjoys singing, joined the campus a capella group The Chorallaries onstage for a spirited rendition of the songs “Arise All Ye of MIT” and “Take Me Back to Tech.” And with that, students filed out of the rink, ready to explore their new home. 

The MIT Quantum Initiative (QMIT) has launched a new postdoctoral fellowship program to accelerate interdisciplinary quantum research and develop the next generation of scientific leaders working at the frontiers of quantum science and technology.

Supported by a grant from the Gordon and Betty Moore Foundation, the program reflects QMIT’s vision of expanding the boundaries of quantum science by encouraging researchers to connect quantum approaches with other disciplines and emerging applications. 

As opportunities in quantum research expand, investing in outstanding early-career researchers has never been more important. These fellowships are designed to help cultivate the next generation of quantum leaders, providing the resources and collaborative environment needed to advance transformative research at MIT.

“Quantum science and technology is in a period of extraordinary opportunity, opening new pathways to solving problems across computation, materials, sensing, and communication. Programs like this help MIT attract outstanding researchers whose ideas will shape the future of the field,” says Anantha Chandrakasan, MIT provost and the Vannevar Bush Professor of Electrical Engineering and Computer Science. 

Launched in December 2025 as an MIT strategic initiative, QMIT brings together researchers from across the Institute to accelerate quantum discovery and apply quantum advances to some of society’s most consequential scientific, technological, industrial, and national security challenges. 

“Quantum science is becoming increasingly interdisciplinary,” says Danna Freedman, the Frederick George Keyes Professor of Chemistry and faculty director of QMIT. “Some of the most exciting breakthroughs will come from researchers who combine deep expertise in quantum with new perspectives from other fields. This fellowship is designed to create exactly those kinds of opportunities.”

The QMIT Fellowship is intentionally designed to foster an interdisciplinary research community. Eligible applicants are outstanding quantum researchers working in a range of fields across physics, chemistry and materials science, and fundamental aspects of biological and Earth sciences. The program specifically seeks researchers whose work combines deep expertise in quantum science with a willingness to explore new intellectual frontiers.

One example of the interdisciplinary vision behind the program is the possibility of applying quantum systems to better understand biological processes, bringing together expertise in atomic physics, quantum algorithms, and biology. The fellows will be embedded across the research areas that define QMIT, including quantum computing, quantum sensing and precision measurement, quantum materials, quantum simulation, and quantum networks. Their research may also explore emerging interdisciplinary approaches that combine artificial intelligence and quantum science.

Fellows supported through the program will join MIT’s extensive quantum ecosystem, working alongside researchers across the Institute, including those affiliated with the Research Laboratory of Electronics, MIT Lincoln Laboratory, the Department of Physics, the Department of Electrical Engineering and Computer Science, the MIT-Harvard Center for Ultracold Atoms, and numerous interdisciplinary research centers and laboratories.

Beyond supporting individual research projects, the fellowship program is intended to strengthen the broader quantum community at MIT by fostering collaboration, mentorship, and intellectual exchange across disciplines.

“Quantum research, in the next few years and across a wide range of domains, is going to make the impossible possible,” says Ian Waitz, MIT’s vice president for research and the head of QMIT. “The QMIT fellowship program is an investment in outstanding postdoctoral scholars who will help bring tremendous new quantum capabilities to unforeseen, creative, and transformative applications in science and technology.”

The inaugural QMIT Fellows will begin their appointments during the 2026 academic year. QMIT expects to open applications for a new cohort in fall 2026 as it continues building a community of researchers working across disciplines to advance the future of quantum science.

When exploring solar system bodies for signs of past or present life, scientists have mainly focused on planets that have (or had) a liquid surface similar to Earth’s. However, mounting evidence suggests that the ingredients for life may exist in a very different environment: the highly acidic clouds that blanket Venus.

Those clouds are made up of about 98 percent sulfuric acid, which scientists had believed to be too acidic for complex biological molecules to survive. But in a new study, MIT researchers have shown that short peptides can not only remain stable in these extremely acidic conditions, they can also fold into shapes that may allow them to have biological functions.

“If peptides find their way to that cloud layer of concentrated sulfuric acid, they will stay and be stably preserved in that cloud of droplets. And once these macromolecules have a defined three-dimensional structure, they can potentially have a function,” says Mei Hong, an MIT professor of chemistry and one of the senior authors of the new study.

The findings suggest that scientists should not rule out planets that don’t resemble Earth in their search for life, says Sara Seager, the Class of 1941 Professor of Planetary Sciences in the Department of Earth, Atmospheric and Planetary Sciences and a professor in the departments of Physics and of Aeronautics and Astronautics.

“We really don’t know the full extent of what planet archetypes are out there. We’re seeking exoplanets that might be a true Earth twin, but what if they’re all Venuses? Our findings definitely open up a whole range of possibilities,” says Seager, another senior author of the study. She will be joining the University of Toronto faculty in September.

Janusz Petkowski, a research assistant professor at Wroclaw University of Science and Technology, is also a senior author of the paper, which appears this week in the Proceedings of the National Academy of Sciences. Jia Yi Zhang, an MIT graduate student, is the paper’s lead author, and former MIT postdoc Aurelio Dregni is also an author. 

Surviving harsh conditions

While Venus’s surface is too hot to be hospitable to life, its cloud layer, which extends from 30 to 40 miles above the planet’s surface, features milder temperatures suitable for life. The clouds are made from droplets of sulfuric acid, which can dissolve metals and destroys most biological molecules on Earth.

Meteorites that contain peptide building blocks regularly enter Venus’s atmosphere, raising the possibility that those peptides could serve as building blocks for simple life forms — if they could survive the clouds’ corrosive environment.

In 2020, Seager’s lab began a series of studies looking at whether different types of biological molecules could persist under those highly acidic conditions. In their initial experiments, working with MIT’s Department of Chemistry Instrumentation Facility (DCIF), they used nuclear magnetic resonance (NMR) spectroscopy — which measures the magnetic properties of atomic nuclei within molecules — to analyze the structures of a variety of molecules in a solution of nearly pure sulfuric acid. 

Those studies showed that nucleic acids, the building blocks of DNA, could remain intact under highly acidic conditions, as could lipids and amino acids. The next step was to figure out if peptides — short strings of amino acids — could persist, and more importantly, whether they could then fold into shapes that might give them biological functions.

For that challenging task, researchers at DCIF suggested that Seager join forces with Hong, an NMR expert who has an advanced 800-megahertz solution NMR spectrometer in her lab. 

To their surprise, the researchers found that the peptides they studied remained stable for many weeks. They believe this is a result of the lack of water in such highly acidic solutions. At 98 percent sulfuric acid, there are very few water molecules, which means that hydrolysis, the chemical reaction that breaks peptide bonds in acid, can’t happen.

“Without water, an acid that you would consider a harsh solvent suddenly is not as menacing as one might think,” Hong says.

After confirming that the peptides remained intact, the researchers began to explore their structures. One of the peptides that the researchers analyzed, a molecule known as HHQ, is a synthetic seven-amino-acid peptide that Hong had previously studied for its role in forming catalytic amyloid fibrils. 

In water, this peptide forms flat beta sheets that eventually form long fibrils. However, in concentrated sulfuric acid, the researchers found that it takes on an entirely different shape — a loop shaped like the Greek letter omega. Such so-called omega loops are occasionally found in some naturally occurring proteins, where they form links between other structural motifs such as sheets or helices.

The other two peptides that the researchers analyzed were a longer variation of HHQ, called HHQ13, and a completely different peptide called K7, which contains seven amino acids. These peptides also formed omega loops in sulfuric acid.

The researchers believe that molecules of sulfuric acid act as a scaffold for the loops, sliding into the center of each loop and holding it in that shape. 

“What hadn’t been known is that peptides can survive so well and have specific three-dimensional shapes in an acidic environment,” Hong says.

Structure and function

In naturally occurring proteins in aqueous solution, omega loops are thought to play a role in protein folding and molecular recognition. Whether they could have other biological functions is not known. However, the fact that peptides can form well-defined, folded structures in acidic environments is an important step in showing that peptides may be able to perform biological functions in such environments.

“Life needs to have specially shaped proteins so that they have a specific target they can latch onto and perform their function. Before this, people thought that peptides couldn’t survive in sulfuric acid, so showing peptides are not only stable, but also fold, is a really big deal,” says Seager, who is leading the Morning Star Missions to Venus.

Adriaan Bax, chief of the Section on Biophysical NMR at the Laboratory of Chemical Physics at the National Institute of Diabetes and Digestive and Kidney Diseases, described the results as “important and unexpected.”

“The observation that these peptides retain a substantial degree of conformational order in concentrated sulfuric acid raises the prospect that folded oligopeptide/protein structures can exist in such environments, potentially supporting the possibility of life in atmospheric conditions that are very different from Earth,” says Bax, who was not involved in the research.

Seager now hopes to pursue additional studies of a molecule called peptide nucleic acid (PNA) — an artificially synthesized molecule that is similar to DNA but with the sugar-phosphate backbone replaced by a peptide backbone. Her lab has previously shown that this molecule, which doesn’t naturally exist on Earth but could offer a potential alternative to DNA, is stable as a single strand in highly acidic environments. She now hopes to study the stability of double-stranded PNA.

The researchers also hope to analyze longer peptides to see if they also take on omega loop shapes, or other structures, in highly concentrated sulfuric acid.

The research was funded by the Alfred P. Sloan Foundation, the NOMIS Foundation, and the National Institutes of Health. 

Sai Ravela, principal research scientist in MIT’s Department of Earth, Atmospheric and Planetary Sciences (EAPS), works with a team of researchers, local partners, and community collaborators to develop game-based computer models to help local communities find solutions to their unique geographical and environmental challenges.

Ravela came to MIT as a postdoc in 2002. Prior to that, he had been working on robotics and computer vision, but he was excited by the idea of studying the climate system and wanted to work in the field of sustainability. “Suddenly, overnight, I became a climate person,” Ravela says.

His project, funded by a 2025 Abdul Latif Jameel Water and Food Systems Lab (J-WAFS) India Grant, explores how agricultural decision-making occurs under climate stress. Using localized climate projections and a participatory approach, the project aims to help communities discover ways to improve their collective agricultural resilience.

EAPS postdoc Anamitra Saha is a key contributor on the grant, working with Ravela and local collaborators to combine downscaled climate modeling, participatory decision-making, and community-based adaptation planning. Other team members include Myisha Ahmad (Carthago Consultancy), Jayanta Basu (University of Calcutta), Anusree Ghosh (Bangladesh Open University), Showmitra Sarkar (Khulna University of Engineering and Technology), and Bivuti Sikder (Dhaka University). 

In a process known as downscaling, researchers take large-scale climate projections and turn them into highly detailed local projections. From these hazard maps, Ravela and Saha can estimate the risk of extreme weather phenomena such as flooding, drought, heat waves, and salinity-related stress. 

“We kind of simulate what the outcome could be in that region,” Ravela explains. “Would it improve agricultural productivity? Would it reduce agricultural productivity? Would it change certain land use patterns? Would the land be less livable, more livable?” 

The team combines surveys, scientific models, and local knowledge to build an impact graph that allows them to explore what might happen to a region during simulated weather events.

Although Ravela knew hazard maps could be useful, he was troubled by how rarely they reached the people whose lives were most affected by the risks they described. “We had clients like insurance companies,” he says. “But I never saw it reach people in a way that made a difference in their lives. And that really bothered me.”

To address this gap, he began thinking about how to help communities engage with hazard maps directly and take part in the decision-making process. In conversations that informed the game’s development, Ravela heard people whose livelihoods are vulnerable to climate events voice immediate concerns about what would happen if a future season failed: “If I don’t plant next season — if I can’t — what would I do?” Ravela wanted to help people think instead about possible choices, different paths, and their respective risks.

When he asked himself what circumstances allow someone to think about risk, the answer began to take shape. “Well, roll a die. Toss a coin,” he thought. “And where do you do these things? In a game.”

How it works

The process the collaborating team developed takes place in three stages. The first is a “snakes and ladders” game, played with physical game pieces and tokens. The second is a mixed game that still uses the gameboard, but a computer generates events and manages portfolios, allowing the system to calculate risk percentages. Once players become comfortable with the mixed game, the final stage, developed by Ravela, abandons the board game and moves fully into a more detailed computer simulation that can be played on a cellphone app.

“We tried this in different stages in three places,” says Ravela. Two villages, Bally Island and Joygopalpur, are in India's Sundarbans region. The third is a village in Bangladesh just across the border. In each location, the work depends on collaboration with local residents, community organizers, and regional partners who help shape the game around local land, water, livelihood, and governance conditions. During development, informal community-engagement sessions helped the team refine and adapt the game. Those interactions also led to intriguing observations that are now helping the team formulate hypotheses for future formal research.

The three villages lie in a coastal region that faces numerous extreme weather events threatening water availability and agricultural productivity. As riverbeds rise from sediment accumulation over time and the land sinks from groundwater extraction, saltwater can more easily intrude into groundwater aquifers, while freshwater drainage, recharge, and flushing become increasingly difficult, intensifying waterlogging and drought. 

“There’s a vicious cycle that’s happening with salinization of the soil,” Ravela explains. 

One visible result is that Boro rice leaves now often begin browning far too early in the season, as salinity and water stress damage crops before they can mature. This cycle occurs in many coastal communities, suggesting to Ravela that the outcomes of the J-WAFS project could have applications around the world.

That broader potential comes from what the game is able to reveal. Instead of treating potential interventions — such as embankments, canals, recharge, crops, fisheries, and energy — as separate choices, the simulation lets players see how each intervention affects the coupled system of land, water, salinity, and livelihoods. When players test different options, simply raising embankments often proves less effective than expected, because it does not break the underlying cycle that causes the land to flood. 

More-integrated strategies — combining mangrove restoration, canal excavation, groundwater recharge, diversified agriculture and fisheries, better water management, and merging solar panels into farming with agrivoltaics or aquavoltaics — can generate better long-term returns while also making the landscape more resilient.

The game also creates space to consider dramatic alternatives to embankment-based protection, including seasonal migration, livelihood shifts, and other difficult choices. These possibilities can be explored safely inside the game, even when they would be almost unimaginable to raise in real life. In this way, difficult questions that might otherwise be avoided can be explored, rather than ignored. And if the game reveals that a difficult choice could lead to better long-term outcomes, that result is not a prescription, but a basis for informed conversation between the community, government, and other decision-makers.

Competition or cooperation?

To make the game effective at developing strategies, Ravela’s team had to understand how many people should play at one time. Too few players may not generate enough diversity of ideas, while too many can slow the process significantly. During game development, groups of roughly ten to twelve people seemed especially workable: large enough to support active interaction, but small enough for practical discussion and learning. 

“Once it crosses a dozen people,” Ravela explains, “it becomes very, very viable as a way to solve problems.”

The games have sparked interest and generated new strategies. People are often excited by the prospect of playing, and repeated play reveals different kinds of expertise. Some participants become especially engaged strategy-explorers; others contribute through discussion, critique, memory, and local knowledge. Together, the process helps identify players who are especially adept at thinking across different dimensions of the problem.

Ravela emphasizes the social aspect of the games as central to their efficacy. “Even though the game is on a phone,” he says, “players are within each other’s reach.” An emcee or facilitator encourages players to engage with one another by asking them to explain their gameplay, discuss their reasoning, and learn from one another’s choices.

While competition is not an explicit feature of the game, there can be zero-sum outcomes. One household’s decision about land, water, drainage, or energy may improve its own outcome while making conditions worse for others. Initially, players may aim for individual success. As they explore longer simulated time horizons, they often shift toward cooperative strategies. 

After each game, the research team and local facilitators lead an educational session where people can learn from each other’s strategies. At first, players often attempt to copy the previous winner’s gameplay — usually, making as much money as possible and saving it in case of disaster. But some disasters are too large for one person to handle alone. 

“That strategy is only optimal up to a certain horizon,” Ravela explains, “because when everyone replicates that strategy, the community doesn’t necessarily thrive.”

As players recognize this, they begin to evolve collective modes of behavior, such as creating a common insurance pool where everyone contributes money to a disaster relief fund. Through multiple iterations of the game, players often appeared to converge on cooperative solutions. 

“The community in this way, playing a game against nature, simulated nature, comes upon solutions that work for them,” says Ravela. “We would love to formally explore this in the future,” Ravela adds.

Why the game works

Ravela’s team sees three advantages to game-based decision-making. First, the game brings new perspectives to the table that formal decision-making often misses. Many communities have strong hierarchies that can discourage women or less powerful community members from participating openly. The game allows people to offer insight without necessarily violating cultural norms. One recurring impression was that women — often responsible for managing family affairs — diversified their portfolios earlier, while men more often concentrated on a single livelihood strategy. The observation was striking enough that the team hopes to test and quantify it formally in future studies.

Second, in the game, all players begin on a level playing field, regardless of status, gender, or wealth. “It democratizes the process,” explains Ravela. In the simulation, a wealthy, influential community figure has no intrinsic advantage over a seamstress. The game reduces natural biases by giving everyone’s ideas a chance to be tested under the same conditions.

Third, because the game is a simulation, people can explore choices that might be too risky, too expensive, or too socially difficult to consider in real life. People may not want to discuss a large aquifer management system, a new land-use arrangement, or a difficult livelihood transition if the real-world implications feel too overwhelming. But inside the game, they can test possibilities without immediate consequence. “So, what, you lose? You start again,” says Ravela.

This is where the game becomes more than a communication tool. It turns uncertainty into a shared decision space. Players can test interventions, observe trade-offs, compare outcomes, and discover strategies before real disasters force those choices upon them. The game shifts the conversation from avoiding risk to reasoning about it, and from fatalistic thinking to collective agency.

Ravela and his collaborators also see the games as a way to address roadblocks in policy implementation by allowing community members to own the solutions they discover. Traditionally, donors may give money to a nongovernmental organization (NGO) that has proposed a project, and the NGO then distributes resources in the community. But it is not always obvious what has actually been implemented, or whether the community has had meaningful ownership of the decision. “In seeking solutions to problems, often the difficulty is developing the policy that provides metrics for the effectiveness of those solutions,” Ravela says. “Games enable people to quickly see the policy space, rather than approaching problems only reactively.”

When people test policies in the game, see how they work, and revise them through repeated play and refinement, they can begin to propose those policies themselves. The result is not simply a technical recommendation from outside experts, but a community-informed basis for action.

What's next?

The broader project, developed with collaborators and community partners in India and Bangladesh, has attracted interest in Bangladesh and Thailand, where similar game-based coastal agricultural resilience projects are being explored. Some customization is necessary to adjust the game to local conditions, but the simulations are highly adaptable. Between 75 and 80 percent of the game can remain the same across locations, while the rest can be tuned to local geography, livelihoods, hazards, and governance structures. Although each place brings its own challenges, “the way land and water and people interact is very similar,” says Ravela.

Building on insights from these game-development and informal community-engagement sessions, Ravela hopes the project can eventually expand to other locations, including members of the Association of Southeast Asian Nations and some places in Latin America. But he emphasizes the importance of establishing longitudinal outcomes before scaling. “The critical question is, does it answer real problems?” he says.

Future formal research will test these emerging hypotheses prospectively and longitudinally. The resulting evidence will help determine whether, where, and how to scale the approach.

If computationally assisted decision-making proves useful over time, the impact could spread far beyond the initial development locations. But the work is not only about finding an optimal solution. It is also about helping people work with one another. As Ravela puts it, “the process really is about helping the people work with each other as much as it is about finding an optimal solution, because part of finding the optimal solution is finding people to work with each other.”

It all started with some exasperated emails. Back in 2009, a group of researchers began venting their frustration with the programming languages designed to help scientists and other researchers perform complex mathematical operations and statistical simulations without learning how to code. These programming languages were rigid and slow. If scientists built something that really worked, they’d need to rewrite the entire program in another language just to run it more quickly.

The emails turned into a research project at MIT with the mission of building an easy-to-use, high-performance programming language called Julia, which is designed for scientific research, data analysis, and modeling complex systems such as jet engines, drugs, financial markets, and robots, to name a few examples.

That research project turned into a lab at MIT, and the lab turned into the company JuliaHub. Along the way, Julia gained a loyal following among scientists, engineers, mathematicians, and others. Today, the free and open-source language counts more than 1 million users, including people working in thousands of companies and universities around the world.

It is only a slight exaggeration to say Julia has been used to model everything under the sun, from the behavior of tiny atoms to semiconductors, neural networks, race cars, and airplanes. It has also been used to study much beyond the sun, with astronomers using Julia for imaging black holes.

Julia’s secret sauce is in the way it compiles code depending on the type of data being used. Such “just-in-time compilation” makes Julia faster and more flexible than other numerical programming languages.

“Scientists and engineers are not programmers. Building scientific applications with multidisciplinary teams of scientists, engineers, and programmers is challenging,” JuliaHub co-founder and CEO Viral Shah says. “We asked: What if you could equip the scientists and engineers with a programming language that allowed them to express their ideas at a high level and also get great software performance?”

Making programming easy for non-programmers has been a north star for JuliaHub’s founders, who include Julia co-creators Shah, MIT professor of mathematics Alan Edelman, Jeff Bezanson SM ’12, PhD ’15, and former MIT research scientist Stefan Karpinski.

In April, JuliaHub’s team took another big step in that direction with the launch of Dyad 3.0, the latest version of its AI platform to help engineering teams accelerate the development of complex physical systems like rockets, heat pumps, and satellites. Engineers are already using Dyad to direct autonomous AI agents as they work through physics simulations, safety analyses, quality controls, and more.

“With Dyad 3.0, you can upload data and design documents and the system will design an entire aircraft for you,” Shah says. “Working with customers like Boeing, we are building agentic hardware design capabilities for engineers. Simplistically, you want to say, ‘Okay computer, build me a plane’; upload the design documents; and have the system account for all the physics, compile all the code, verify everything, and build the entire design agentically.”

Humble beginnings

After discussing the need for better programming languages for scientists and other researchers, Julia’s co-creators started the Julia Lab around 2009. The Julia Lab remains active in MIT’s Computer Science and Artificial Intelligence Laboratory.

The core idea was to create a high-performance platform that would excel at engineering, scientific, and mathematics applications. Shah says before Julia, scientists and engineers would either have to hire someone to build software for them or accept the slow performance of the few programming languages designed for them.

“We wanted to create something as easy to use as Python or MATLAB but as fast as the C programming language,” Shah says. “We built Julia for ourselves.”

Edelman says at first, the researchers didn’t think anyone would want their creation.

“We figured it would take 10 years before anyone was interested, but we said, ‘Patience is a virtue, so let’s do it,’” Edelman recalls.

The MIT researchers announced Julia with a blog post in 2012. They quickly realized many other researchers shared their frustration.

“When we first started, we were targeting interactive research workflows, but increasingly people are using it for everything,” Bezanson says. “Now we’re moving the whole stack of the language onto smaller, embedded devices as we evolve with our users.”

Since those early days, Edelman has taught a class on Julia with students from nearly every department at MIT. Today, he often learns students are already using Julia when they enroll in the class for applications as wide ranging as robotics, astronomy, physics simulations, and finance.

“Researchers come up to me and say, ‘I tell my supervisor I’m using Julia because it’s fast, but don’t tell them I’m using Julia because it’s really fun,’” Edelman says. “The key thing is Julia’s abstractions. A lot of times a coding language forces you to solve the one problem you’re thinking about. Julia’s language makes it so you’re solving not only the problem you’re thinking about, but other people’s problems around the world too. It encourages you to solve problems more generally.”

As Julia gained popularity, researchers around the world started asking the Julia team for support. By 2015, the demand became strong enough that they decided to start JuliaHub and help users through the company full-time. They received support from the MIT Deshpande Center for Technological Innovation and others at MIT to get the company off the ground.

JuliaHub’s work has evolved from simply helping users to advancing the language more generally. That’s powered an impressive list of creations from Julia’s loyal users. Julia has been used to simulate computer circuits, detect health disparities, model global climates and oceans, analyze brain activity, and more. 

After someone built a pharmaceutical modeling platform in Julia, it was used to accelerate development of Moderna’s Covid-19 vaccine. In another case, researchers used Julia to create a program for avoiding aircraft collisions. They found it ran about 50 times faster than an earlier version built on Python. Engineers at Meta used Julia to develop a better audio codec for WhatsApp’s 4 billion users.

“Over the years we’ve seen industrial, government, and academic users doing all kinds of interesting things with the Julia language,” Edelman says. “It’s honestly surprised us in many ways, the wide-ranging things people are using it for.”

Autonomous design

JuliaHub launched Dyad 1.0 in June of 2025 as a research agent to accelerate programming and Dyad 2.0 in December. The founders believe Dyad 3.0 represents a new level of ability and autonomy for designing complex systems.

“One important thing about Dyad is that it is a physics compiler and hence enforces physical laws,” Shah explains. “General AI systems often solve physical problems in ways that violate physical laws. When using the Dyad agent, it will detect such violations and guide the agent in the direction of the physically correct solution. We expect it will decrease design times in product engineering by orders of magnitude, leading to months of work being accomplished in hours.”

One way Edelman sees the impact of Julia is through his class. One student recently used Dyad to model how robots move around in space. Another used it to build a rocket engine.

“At the end he said, ‘I couldn’t believe how easy that was — I just got a rocket engine!’” Edelman recalls.

This past spring, MIT physics graduate student Manu Srivastava opened an email from a group of high school students in India he had never met.

They were hoping to enter Beamline for Schools, an international competition that gives secondary school students the chance to design and carry out experiments using particle accelerator beams. And they were looking for a mentor.

Srivastava, who studies quantum gravity as a PhD student in the MIT Center for Theoretical Physics – a Leinweber Institute, with Professor Hong Liu, gets other requests to mentor students, often through companies charging families for access to scientists or students at prestigious universities. He usually declines, but this message came directly from the students.

“I've also cold-emailed a lot in my early career, and it usually never works,” he says. “But this email seemed very genuine. They wanted to do something nice and they just needed some guidance.”

Many months and many more emails and calls later, the students secured a place with Srivastava to attend CERN, in Geneva, where they spent two weeks turning their proposed idea into a real experiment. 

Finding an experiment worth doing

Calling themselves Team attoPION, the students are one of five teams selected in the 13th annual Beamline for Schools competition from a record 712 teams representing 89 countries and more than 4,500 students. The six high schoolers met through a combination of science competitions and mutual friends, and attend four schools in four cities across India.

When they first met with Srivastava, the students already had several experimental ideas. His role, he says, was to help determine which directions were practical and scientifically interesting.

They settled on measuring pion charge exchange. Pions are short-lived subatomic particles that can carry positive, negative, or neutral charge. In the process the students want to study, a positively charged pion interacts with a neutron in a target material, producing a neutral pion and a positively charged proton. The team wants to characterize how often that reaction occurs.

Srivastava suspected such a measurement could have relevance to the Deep Underground Neutrino Experiment, or DUNE, a major international experiment designed to study neutrinos.

Dave Newbold, a co-spokesperson for DUNE, says understanding how pions interact with matter helps researchers quantify uncertainties in DUNE’s measurements. In particular, pion interactions can affect estimates of a neutrino’s flavor and energy, which researchers need to measure accurately to determine whether they have observed something new.

And although Beamline for Schools has an educational mission, Newbold says the students aren't simply reproducing a classroom demonstration. “The proposal is real experimental particle physics!” he notes.

If successful, Newbold believes the work could improve scientists' understanding of this particular interaction and potentially lead to a publishable result. Similar “test beam” experiments remain important tools in particle physics: DUNE's detector designs were themselves demonstrated using the (albeit much larger) ProtoDUNE experiments at CERN.

“This [proposal] stands out because of the work the students have put into motivating their measurement, and demonstrating that the experiment is feasible,” Newbold says. “It's certainly at a level far above anything I was thinking about at high school.”

Learning to navigate uncertainty

At CERN, the students worked hands-on with detectors and data-acquisition systems, collected and analyze data, and attended talks by CERN scientists.

In advance of the trip, the team worked with Berare Göktürk, one of the support scientists for Beamline for Schools. In their preparation sessions for the experiment, they realized that the charge-exchange process they hope to observe is extremely rare, forcing them to think through how they might reliably detect it.

With just a few months months to prepare and only 12 days of test-beam time, Göktürk cautioned that producing a result useful to a much larger experiment would be an ambitious outcome.

“We prepare in the best way possible, but we also stay humble and we are aware of the limitations we have,” she says. Her priority is for the students to “understand the journey of a scientist” as they encounter technical problems and work together to solve them.

For Srivastava, mentoring an experiment has also taken him well outside his own specialty. A theoretical physicist, he credits MIT's culture with encouraging him to follow questions beyond the boundaries of his research, including by attending seminars, colloquia, and research meetings across physics.

The experience has been personally meaningful for Srivastava, who grew up in India and sees the mentorship as a way to encourage young people there to pursue fundamental science. 

“I didn't even know what CERN was in high school,” he says. “But these students, they are just that good. They deserve all the credit.”

Although Jacob Becraft had two swings and two misses when he first tried to become an entrepreneur as a graduate student, the MIT Sandbox Innovation Fund Program allowed him to keep at it. This especially benefited cancer patients, as Becraft went on to co-found Strand Therapeutics: a $550-million firm whose programmable mRNA drug has shrunk tumors in patients who had exhausted all other treatment options.

Stories like Becraft’s took center stage at the recent 10-year anniversary celebration of the MIT Sandbox Innovation Fund Program, where student founders, alumni, mentors, and university leaders gathered to reflect on a decade of empowering student entrepreneurs. Speaking at the event, Becraft referred to Strand as "our third swing at the plate," explaining that the Sandbox model gave him "the freedom and ability to fail fast" — letting previous venture ideas "blow up in our faces" before moving on.

For Strand, Becraft says, MIT Sandbox helped him and his co-founder, Tasuku Kitada, to "get out, do some travel, some market research, meet with experts in the field, meet with mentors who could help us build the company — and eventually find investors who were going to back this big vision to transform medicine."

MIT Sandbox was launched in 2016 by Ian Waitz, then-dean of the School of Engineering and now MIT's vice president for research, to lower the barrier for students to try entrepreneurship. The concept of a new program focused on student-led entrepreneurship was developed in consultation with internal MIT leaders and supporters of MIT, including Alan Spoon, a life member emeritus of the MIT Corporation. From its inception, MIT Sandbox has been open to all MIT students, from undergraduates to PhD students. Teams are awarded between $500 and $5,000 to begin their process, and they are matched with two mentors and connected with other expert advisors. 

As they make progress, students can go before the program’s funding board to ask for up to $25,000. Supported entirely by alumni, corporate sponsors, entrepreneurs, and investors, the program has grown to include about 350 teams each semester, some of which are new and some continuing their participation according to their own timelines.

Anantha P. Chandrakasan, MIT provost, explained in the program's decade-in-review report: "Since its inception 10 years ago, MIT Sandbox has been a defining part of MIT's innovation ecosystem, ensuring that every student with the curiosity to explore entrepreneurship has the resources, mentorship, and community to take their first steps."

MIT Sandbox is a "home," where students can "explore, seriously test assumptions, talk to customers, build prototypes, fail, pivot, learn, and grow," says Jinane Abounadi, founding executive director of Sandbox. "And they can do that with a lot of support — and I don't just mean financial support. I mean a lot of support from a lot of people."

For Samuel Udotong, co-founder and CTO of Fireflies.ai, early funding was the difference between an idea and a company. "I think largely because we had gotten a little bit of Sandbox funding, we were actually able to take the risk to move out to San Francisco and try to build the company," he says. "But it would have been really a money barrier if we hadn't gotten the initial $5,000 from Sandbox."

Startup investor and advisor Sophie V. Vandebroek says, "MIT has extraordinary students from around the globe as well as faculty who are top experts in their fields. What’s often lacking," she says, "is confidence. That is where Sandbox plays a vital role. Sandbox enables every individual student to believe that they can be an entrepreneur."

At the anniversary celebration, Fred Parietti, co-founder and CEO of Multiply Labs, recounted how his early product prototypes were developed on his kitchen table and had to be moved regularly according to the dictates of his grad school housemates. Those prototypes wouldn't have been built at all, he said, without MIT Sandbox.

The first funding he received was minimal, "but it wasn't zero, and zero represented my resources as a student. That belief in us and the possibility to build a prototype were game-changers," Parietti said.

Multiply Labs, with 60-plus employees, has raised $36 million and develops robotics technology to manufacture biological drugs safely and economically. The firm supplies pharmaceutical customers including AstraZeneca and Kyverna Therapeutics, whose chief medical and development officer, Naji Gehchan, is an MIT Sandbox mentor.

That same willingness to back an unconventional approach helped AeroShield get off the ground. "One of the things that enables me to stand here today is that Sandbox created a safe environment where it was encouraged to look at this problem backwards, rather than from the nanostructure up," says Elise Strobach, CEO and founder of AeroShield.

The anniversary celebration speakers also included Ross Finman, CEO and founder of Augmodo; Laureen Meroueh, CEO and founder of Hertha Metals; and Daris Bunadar, chief scientist at Lightmatter. All were working on their PhDs when they started exploring commercial applications of their research. All recognize the critical role that MIT Sandbox, in addition to other programs — such as the MIT I-Corps Program, the Martin Trust Center for MIT Entrepreneurship, MIT Venture Mentoring Service (VMS), and the Bernard M. Gordon-MIT Engineering Leadership Program — played in their development as entrepreneurs. These programs offered the space to explore the possibility of not only founding a deep tech company, but also taking on an executive role as their ventures raised venture capital and grew into substantial companies. Today they all have big ambitions for the growth and impact of their companies — ambitions that are made possible only thanks to innovative technologies and an entrepreneurial drive. 

Over its decade of existence, MIT Sandbox has supported over 4,000 teams, representing 8,000 participants associated with a wide range of industries and nonprofit endeavors. It has disbursed more than $11 million in non-dilutive funding, meaning the program takes no stake in the resulting ventures. MIT Sandbox has been involved in the creation of 475 companies in more than 30 countries, and companies that were started in the program have raised $8.7 billion in venture funding.

MIT Sandbox collaborates with other programs across MIT — including the Martin Trust Center, VMS, Kuo Sharp Center, MITdesignX, the PKG Center for Social Impact, the MIT Climate Project, I-Corps, and others — and its teams have excelled in innovation accelerators and competitions. Nine out of 10 winners of MIT's $100K Entrepreneurship Competition have been MIT Sandbox participants.

Apart from the program's impressive results, MIT Sandbox aims to first and foremost serve as a great educational tool, developing the innovators themselves.

"From an educator's perspective, this is just another incredible way to teach," said Abounadi at the anniversary celebration. "MIT Sandbox is a place where students can start seeing themselves as people who can create a meaningful impact in the world," she said, "and that is really what innovation and entrepreneurship are all about."

Paula T. Hammond, School of Engineering dean and Institute Professor, echoed the same sentiments: "What I find most compelling, year after year, is not only what students build, but how they change. They gain confidence, learn to refine before they scale, and begin to see themselves as people who can create meaningful impact, strengthening not only their own trajectories, but the broader MIT community."