In the Northeastern United States, the Gulf of Maine represents one of the most biologically diverse marine ecosystems on the planet — home to whales, sharks, jellyfish, herring, plankton, and hundreds of other species. But even as this ecosystem supports rich biodiversity, it is undergoing rapid environmental change. The Gulf of Maine is warming faster than 99 percent of the world’s oceans, with consequences that are still unfolding.

A new research initiative developing at MIT Sea Grant, called LOBSTgER — short for Learning Oceanic Bioecological Systems Through Generative Representations — brings together artificial intelligence and underwater photography to document the ocean life left vulnerable to these changes and share them with the public in new visual ways. Co-led by underwater photographer and visiting artist at MIT Sea Grant Keith Ellenbogen and MIT mechanical engineering PhD student Andreas Mentzelopoulos, the project explores how generative AI can expand scientific storytelling by building on field-based photographic data.

Just as the 19th-century camera transformed our ability to document and reveal the natural world — capturing life with unprecedented detail and bringing distant or hidden environments into view — generative AI marks a new frontier in visual storytelling. Like early photography, AI opens a creative and conceptual space, challenging how we define authenticity and how we communicate scientific and artistic perspectives. 

In the LOBSTgER project, generative models are trained exclusively on a curated library of Ellenbogen’s original underwater photographs — each image crafted with artistic intent, technical precision, accurate species identification, and clear geographic context. By building a high-quality dataset grounded in real-world observations, the project ensures that the resulting imagery maintains both visual integrity and ecological relevance. In addition, LOBSTgER’s models are built using custom code developed by Mentzelopoulos to protect the process and outputs from any potential biases from external data or models. LOBSTgER’s generative AI builds upon real photography, expanding the researchers’ visual vocabulary to deepen the public’s connection to the natural world.

At its heart, LOBSTgER operates at the intersection of art, science, and technology. The project draws from the visual language of photography, the observational rigor of marine science, and the computational power of generative AI. By uniting these disciplines, the team is not only developing new ways to visualize ocean life — they are also reimagining how environmental stories can be told. This integrative approach makes LOBSTgER both a research tool and a creative experiment — one that reflects MIT’s long-standing tradition of interdisciplinary innovation.

Underwater photography in New England’s coastal waters is notoriously difficult. Limited visibility, swirling sediment, bubbles, and the unpredictable movement of marine life all pose constant challenges. For the past several years, Ellenbogen has navigated these challenges and is building a comprehensive record of the region’s biodiversity through the project, Space to Sea: Visualizing New England’s Ocean Wilderness. This large dataset of underwater images provides the foundation for training LOBSTgER’s generative AI models. The images span diverse angles, lighting conditions, and animal behaviors, resulting in a visual archive that is both artistically striking and biologically accurate.

LOBSTgER’s custom diffusion models are trained to replicate not only the biodiversity Ellenbogen documents, but also the artistic style he uses to capture it. By learning from thousands of real underwater images, the models internalize fine-grained details such as natural lighting gradients, species-specific coloration, and even the atmospheric texture created by suspended particles and refracted sunlight. The result is imagery that not only appears visually accurate, but also feels immersive and moving.

The models can both generate new, synthetic, but scientifically accurate images unconditionally (i.e., requiring no user input/guidance), and enhance real photographs conditionally (i.e., image-to-image generation). By integrating AI into the photographic workflow, Ellenbogen will be able to use these tools to recover detail in turbid water, adjust lighting to emphasize key subjects, or even simulate scenes that would be nearly impossible to capture in the field. The team also believes this approach may benefit other underwater photographers and image editors facing similar challenges. This hybrid method is designed to accelerate the curation process and enable storytellers to construct a more complete and coherent visual narrative of life beneath the surface.

In one key series, Ellenbogen captured high-resolution images of lion’s mane jellyfish, blue sharks, American lobsters, and ocean sunfish (Mola mola) while free diving in coastal waters. “Getting a high-quality dataset is not easy,” Ellenbogen says. “It requires multiple dives, missed opportunities, and unpredictable conditions. But these challenges are part of what makes underwater documentation both difficult and rewarding.”

Mentzelopoulos has developed original code to train a family of latent diffusion models for LOBSTgER grounded on Ellenbogen’s images. Developing such models requires a high level of technical expertise, and training models from scratch is a complex process demanding hundreds of hours of computation and meticulous hyperparameter tuning.

The project reflects a parallel process: field documentation through photography and model development through iterative training. Ellenbogen works in the field, capturing rare and fleeting encounters with marine animals; Mentzelopoulos works in the lab, translating those moments into machine-learning contexts that can extend and reinterpret the visual language of the ocean.

“The goal isn’t to replace photography,” Mentzelopoulos says. “It’s to build on and complement it — making the invisible visible, and helping people see environmental complexity in a way that resonates both emotionally and intellectually. Our models aim to capture not just biological realism, but the emotional charge that can drive real-world engagement and action.”

LOBSTgER points to a hybrid future that merges direct observation with technological interpretation. The team’s long-term goal is to develop a comprehensive model that can visualize a wide range of species found in the Gulf of Maine and, eventually, apply similar methods to marine ecosystems around the world.

The researchers suggest that photography and generative AI form a continuum, rather than a conflict. Photography captures what is — the texture, light, and animal behavior during actual encounters — while AI extends that vision beyond what is seen, toward what could be understood, inferred, or imagined based on scientific data and artistic vision. Together, they offer a powerful framework for communicating science through image-making.

In a region where ecosystems are changing rapidly, the act of visualizing becomes more than just documentation. It becomes a tool for awareness, engagement, and, ultimately, conservation. LOBSTgER is still in its infancy, and the team looks forward to sharing more discoveries, images, and insights as the project evolves.

Answer from the lead image: The left image was generated using using LOBSTgER’s unconditional models and the right image is real.

For more information, contact Keith Ellenbogen and Andreas Mentzelopoulos.

Modern fighter jets contain hundreds or even thousands of sensors. Some of those sensors collect data every second, others every nanosecond. For the engineering teams building and testing those jets, all those data points are hugely valuable — if they can make sense of them.

Nominal is an advanced software platform made for engineers building complex systems ranging from fighter jets to nuclear reactors, satellites, rockets, and robots. Nominal’s flagship product, Nominal Core, helps teams organize, visualize, and securely share data from tests and operations. The company’s other product, Nominal Connect, helps engineers build custom applications for automating and syncing their hardware systems.

“It’s a very technically challenging problem to take the types of data that our customers are generating and get them into a single place where people can collaborate and get insights,” says Nominal co-founder Jason Hoch ’13. “It’s hard because you’re dealing with a lot of different data sources, and you want to be able to correlate those sources and apply mathematical formulas. We do that automatically.”

Hoch started Nominal with Cameron McCord ’13, SM ’14 and Bryce Strauss after the founders had to work with generic data tools or build their own solutions at places like Lockheed Martin and Anduril. Today, Nominal is working with organizations in aerospace, defense, robotics, manufacturing, and energy to accelerate the development of products critical for applications in U.S. national security and beyond.

“We built Nominal to take the best innovations in software and data technology and tailor them to the workflows that engineers go through when building and testing hardware systems,” McCord says. “We want to be the data and software backbone across all of these types of organizations.”

Accelerating hardware development

Hoch and McCord met during their first week at MIT and joined the same fraternity as undergraduates. Hoch double majored in mathematics and computer science and engineering, and McCord participated in the Navy Reserve Officers’ Training Corps (NROTC) while majoring in physics and nuclear science and engineering.

“MIT let me flex my technical skills, but I was also interested in the broader implications of technology and national security,” McCord says. “It was an interesting balance where I was learning the hardcore engineering skills, but always having a wider aperture to understand how the technology I was learning about was going to impact the world.”

Following MIT, McCord spent eight years in the Navy before working at the defense technology company Anduril, where he was charged with building the software systems to test different products. Hoch also worked at the intelligence and defense-oriented software company Palantir.

McCord met Strauss, who had worked as an engineer at Lockheed Martin, while the two were at Harvard Business School. The eventual co-founders realized they had each struggled with software during complex hardware development projects, and set out to build the tools they wished they’d had.

At the heart of Nominal’s platform is a unified database that can connect and organize hundreds of data sources in real-time. Nominal’s system allows engineers to search through or visualize that information, helping them spot trends, catch critical events, and investigate anomalies — what Nominal’s team describes as learning the rules governing complex systems.

“We’re trying to get answers to engineers so they understand what’s happening and can keep projects moving forward,” says Strauss. “Testing and validating these systems are fundamental bottlenecks for hardware progress. Our platform helps engineers answer questions like, ‘When we made a 30-degree turn at 16,000 feet, what happened to the engine’s temperature, and how does that compare to what happened yesterday?’”

By automating tasks like data stitching and visualization, Nominal’s platform helps accelerate post-test analysis and development processes for complex systems. And because the platform is cloud-hosted, engineers can easily share visualizations and other dynamic assets with members of their team as opposed to making static reports, allowing more people in an organization to interact directly with the data.

From satellites to drones, robots to rockets

Nominal recently announced a $75 million Series B funding round, led by Sequoia Capital, to accelerate their growth.

“We’ll use the funds to accelerate product roadmaps for our existing products, launch new products across the hardware test stack, and more than double our team,” says McCord.

Today, aerospace customers are using Nominal’s platform to monitor their assets in orbit. Manufacturers are using Nominal to make sure their components work as expected before they’re integrated into larger systems. Nuclear fusion companies are using Nominal to understand when their parts might fail due to heat.

“The products we’ve built are transferrable,” Hoch says. “It doesn’t matter if you’re building a nuclear fusion reactor or a satellite, those teams can benefit from the Nominal tool chain.”

Ultimately the founders believe the platform helps create better products by enabling a data-driven, iterative design process more commonly seen in the software development industry.

“The concept of continuous integration and development in software revolutionized the industry 20 years ago. Before that, it was common to build software in large, slow batches – developing for months, then testing and releasing all at once,” Strauss explains. “We’re bringing continuous testing to hardware. It’s about constantly creating that feedback loop to improve performance. It’s a new paradigm for how hardware is built. We’ve seen companies like SpaceX do this well to move faster and outpace the competition. Now, that approach is available to everyone.”

Since MIT opened the first-of-its-kind venture studio within a university in 2019, it has demonstrated how a systemic process can help turn research into impactful ventures. 

Now, MIT Proto Ventures is launching the “R&D Venture Studio Playbook,” a resource to help universities, national labs, and corporate R&D offices establish their own in-house venture studios. The online publication offers a comprehensive framework for building ventures from the ground up within research environments.

“There is a huge opportunity cost to letting great research sit idle,” says Fiona Murray, associate dean for innovation at the MIT Sloan School of Management and a faculty director for Proto Ventures. “The venture studio model makes research systematic, rather than messy and happenstance.” 

Bigger than MIT

The new playbook arrives amid growing national interest in revitalizing the United States’ innovation pipeline — a challenge underscored by the fact that just a fraction of academic patents ever reach commercialization.

“Venture-building across R&D organizations, and especially within academia, has been based on serendipity,” says MIT Professor Dennis Whyte, a faculty director for Proto Ventures who helped develop the playbook. “The goal of R&D venture studios is to take away the aspect of chance — to turn venture-building into a systemic process. And this is something not just MIT needs; all research universities and institutions need it.”

Indeed, MIT Proto Ventures is actively sharing the playbook with peer institutions, federal agencies, and corporate R&D leaders seeking to increase the translational return on their research investments.

“We’ve been following MIT’s Proto Ventures model with the vision of delivering new ventures that possess both strong tech push and strong market pull,” says Mark Arnold, associate vice president of Discovery to Impact and managing director of Texas startups at The University of Texas at Austin. “By focusing on market problems first and creating ventures with a supportive ecosystem around them, universities can accelerate the transition of ideas from the lab into real-world solutions.” 

What’s in the playbook

The playbook outlines the venture studio model process followed by MIT Proto Ventures. MIT’s venture studio embeds full-time entrepreneurial scientists — called venture builders — inside research labs. These builders work shoulder-to-shoulder with faculty and graduate students to scout promising technologies, validate market opportunities, and co-create new ventures.

“We see this as an open-source framework for impact,” says MIT Proto Ventures Managing Director Gene Keselman. “Our goal is not just to build startups out of MIT — it’s to inspire innovation wherever breakthrough science is happening.”

The playbook was developed by the MIT Proto Ventures team — including Keselman, venture builders David Cohen-Tanugi and Andrew Inglis, and faculty leaders Murray, Whyte, Andrew Lo, Michael Cima, and Michael Short. 

“This problem is universal, so we knew if it worked there’d be an opportunity to write the book on how to build a translational engine,” Keselman said. “We’ve had enough success now to be able to say, ‘Yes, this works, and here are the key components.’” 

In addition to detailing core processes, the playbook includes case studies, sample templates, and guidance for institutions seeking to tailor the model to fit their unique advantages. It emphasizes that building successful ventures from R&D requires more than mentorship and IP licensing — it demands deliberate, sustained focus, and a new kind of translational infrastructure. 

How it works

A key part of MIT’s venture studio is structuring efforts into distinct tracks or problem areas — MIT Proto Ventures calls these channels. Venture builders work in a single track that aligns with their expertise and interest. For example, Cohen-Tanugi is embedded in the MIT Plasma Science and Fusion Center, working in the Fusion and Clean Energy channel. His first two venture successes have been a venture using superconducting magnets for in-space propulsion and a deep-tech startup improving power efficiency in data centers.

“This playbook is both a call to action and a blueprint,” says Cohen-Tanugi, lead author of the playbook. “We’ve learned that world-changing inventions often remain on the lab bench not because they lack potential, but because no one is explicitly responsible for turning them into businesses. The R&D venture studio model fixes that.”

Four MIT rising seniors have been selected to receive a 2025 Barry Goldwater Scholarship, including Avani Ahuja and Jacqueline Prawira in the School of Engineering and Julianna Lian and Alex Tang from the School of Science. An estimated 5,000 college sophomores and juniors from across the United States were nominated for the scholarships, of whom only 441 were selected.

The Goldwater Scholarships have been conferred since 1989 by the Barry Goldwater Scholarship and Excellence in Education Foundation. These scholarships have supported undergraduates who go on to become leading scientists, engineers, and mathematicians in their respective fields.

Avani Ahuja, a mechanical engineering and electrical engineering major, conducts research in the Conformable Decoders group, where she is focused on developing a “wearable conformable breast ultrasound patch” that makes ultrasounds for breast cancer more accessible.

“Doing research in the Media Lab has had a huge impact on me, especially in the ways that we think about inclusivity in research,” Ahuja says.

In her research group, Ahuja works under Canan Dagdeviren, the LG Career Development Professor of Media Arts and Sciences. Ahuja plans to pursue a PhD in electrical engineering. She aspires to conduct research in electromechanical systems for women’s health applications and teach at the university level.

“I want to thank Professor Dagdeviren for all her support. It’s an honor to receive this scholarship, and it’s amazing to see that women’s health research is getting recognized in this way,” Ahuja says.

Julianna Lian studies mechanochemistry, organic, and polymer chemistry in the lab of Professor Jeremiah Johnson, the A. Thomas Guertin Professor of Chemistry. In addition to her studies, she serves the MIT community as an emergency medical technician (EMT) with MIT Emergency Medical Services, is a member of MIT THINK, and a ClubChem mentorship chair.

“Receiving this award has been a tremendous opportunity to not only reflect on how much I have learned, but also on the many, many people I have had the chance to learn from,” says Lian. “I am deeply grateful for the guidance, support, and encouragement of these teachers, mentors, and friends. And I am excited to carry forward the lasting curiosity and excitement for chemistry that they have helped inspire in me.”

Lian’s career goals post-graduation include pursuing a PhD in organic chemistry, to conduct research at the interface of synthetic chemistry and materials science, aided by computation, and to teach at the university level.

Jacqueline Prawira, a materials science and engineering major, joined the Center of Decarbonization and Electrification of Industry as a first-year Undergraduate Research Opportunities Program student and became a co-inventor on a patent and a research technician at spinout company Rock Zero. She has also worked in collaboration with Indigenous farmers and Diné College students on the Navajo Nation.

“I’ve become significantly more cognizant of how I listen to people and stories, the tangled messiness of real-world challenges, and the critical skills needed to tackle complex sustainability issues,” Prawira says.

Prawira is mentored by Yet-Ming Chiang, professor of materials science and engineering. Her career goals are to pursue a PhD in materials science and engineering and to research sustainable materials and processes to solve environmental challenges and build a sustainable society.

“Receiving the prestigious title of 2025 Goldwater Scholar validates my current trajectory in innovating sustainable materials and demonstrates my growth as a researcher,” Prawira says. “This award signifies my future impact in building a society where sustainability is the norm, instead of just another option.”

Alex Tang studies the effects of immunotherapy and targeted molecular therapy on the tumor microenvironment in metastatic colorectal cancer patients. He is supervised by professors Jonathan Chen at Northwestern University and Nir Hacohen at the Broad Institute of MIT and Harvard.

“My mentors and collaborators have been instrumental to my growth since I joined the lab as a freshman. I am incredibly grateful for the generous mentorship and support of Professor Hacohen and Professor Chen, who have taught me how to approach scientific investigation with curiosity and rigor,” says Tang. “I’d also like to thank my advisor Professor Adam Martin and first-year advisor Professor Angela Belcher for their guidance throughout my undergraduate career thus far. I am excited to carry forward this work as I progress in my career.” Tang intends to pursue physician-scientist training following graduation.

The Scholarship Program honoring Senator Barry Goldwater was designed to identify, encourage, and financially support outstanding undergraduates interested in pursuing research careers in the sciences, engineering, and mathematics. The Goldwater Scholarship is the preeminent undergraduate award of its type in these fields.

In 2025, MIT granted tenure to 11 faculty members across the School of Engineering. This year’s tenured engineers hold appointments in the departments of Aeronautics and Astronautics, Biological Engineering, Chemical Engineering, Electrical Engineering and Computer Science (EECS) — which reports jointly to the School of Engineering and MIT Schwarzman College of Computing — Materials Science and Engineering, Mechanical Engineering, and Nuclear Science and Engineering.

“It is with great pride that I congratulate the 11 newest tenured faculty members in the School of Engineering. Their dedication to advancing their fields, mentoring future innovators, and contributing to a vibrant academic community is truly inspiring,” says Anantha Chandrakasan, chief innovation and strategy officer, dean of engineering, and the Vannevar Bush Professor of Electrical Engineering and Computer Science who will assume the title of MIT provost July 1. “This milestone is not only a testament to their achievements, but a promise of even greater impact ahead.”

This year’s newly tenured engineering faculty include:

Bryan Bryson, the Phillip and Susan Ragon Career Development Professor in the Department of Biological Engineering, conducts research in infectious diseases and immunoengineering. He is interested in developing new tools to dissect the complex dynamics of bacterial infection at a variety of scales ranging from single cells to infected animals, sitting in both “reference frames” by taking both an immunologist’s and a microbiologist’s perspective.

Connor Coley is the Class of 1957 Career Development Professor and associate professor of chemical engineering, with a shared appointment in EECS. His research group develops new computational methods at the intersection of artificial intelligence and chemistry with relevance to small molecule drug discovery, chemical synthesis, and structure elucidation.

Mohsen Ghaffari is the Steven and Renee Finn Career Development Professor and an associate professor in the EECS. His research explores the theory of distributed and parallel computation. He has done influential work on a range of algorithmic problems, including generic derandomization methods for distributed computing and parallel computing, improved distributed algorithms for graph problems, sublinear algorithms derived via distributed techniques, and algorithmic and impossibility results for massively parallel computation.

Rafael Gomez-Bombarelli, the Paul M. Cook Development Professor and associate professor of materials science and engineering, works at the interface between machine learning and atomistic simulations. He uses computational tools to tackle design of materials in complex combinatorial search spaces, such as organic electronic materials, energy storage polymers and molecules, and heterogeneous (electro)catalysts. 

Song Han, an associate professor in EECS, is a pioneer in model compression and TinyML. He has innovated in key areas of pruning quantization, parallelization, KV cache optimization, long-context learning, and multi-modal representation learning to minimize generative AI costs, and he designed the first hardware accelerator (EIE) to exploit weight sparsity.

Kaiming He, the Douglass Ross (1954) Career Development Professor of Software Technology and an associate professor in EECS, is best known for his work on deep residual networks (ResNets). His research focuses on building computer models that can learn representations and develop intelligence from and for the complex world, with the long-term goal of augmenting human intelligence with more capable artificial intelligence.

Phillip Isola, the Class of 1948 Career Development Professor and associate professor in EECS, studies computer vision, machine learning, and AI. His research aims to uncover fundamental principles of intelligence, with a particular focus on how models and representations of the world can be acquired through self-supervised learning, from raw sensory experience alone, and without the use of labeled data.

Mingda Li is the Class of 1947 Career Development Professor and an associate professor in the Department of Nuclear Science and Engineering. His research lies in characterization and computation.

Richard Linares is an associate professor in the Department of Aeronautics and Astronautics. His research focuses on astrodynamics, space systems, and satellite autonomy. Linares develops advanced computational tools and analytical methods to address challenges associated with space traffic management, space debris mitigation, and space weather modeling.

Jonathan Ragan-Kelley, an associate professor in EECS, has designed everything from tools for visual effects in movies to the Halide programming language that’s widely used in industry for photo editing and processing. His research focuses on high-performance computer graphics and accelerated computing, at the intersection of graphics with programming languages, systems, and architecture.

Arvind Satyanarayan is an associate professor in EECS. His research areas cover data visualization, human-computer interaction, and artificial intelligence and machine learning. He leads the MIT Visualization Group, which uses interactive data visualization as a petri dish to study intelligence augmentation — how computation can help amplify human cognition and creativity while respecting our agency.

Startups and government defense agencies have historically seemed like polar opposites. Startups thrive on speed and risk, while defense agencies are more cautious. Over the past few years, however, things have changed. Many startups are eager to work with these organizations, which are always looking for innovative solutions to their hardest problems.

To help bridge that gap while advancing research along the way, MIT Lecturer Gene Keselman launched MIT’s Mission Innovation X (MIx) along with Sertac Karaman, a professor in the MIT Department of Aeronautics and Astronautics, and Fiona Murray, the William Porter Professor of Entrepreneurship at the MIT Sloan School of Management. MIx develops educational programming, supports research at MIT, and facilitates connections among government organizations, startups, and researchers.

“Startups know how to commercialize their tech, but they don’t necessarily know how to work with the government, and especially how to understand the needs of defense customers,” explains MIx Senior Program Manager Keenan Blatt. “There are a lot of different challenges when it comes to engaging with defense, not only from a procurement cycle and timeline perspective, but also from a culture perspective.”

MIx’s work helps innovators secure crucial early funding while giving defense agencies access to cutting-edge technologies, boosting America’s security capabilities in the process. Through the work, MIx has also become a thought leader in the emerging “dual-use” space, in which researchers and founders make strategic choices to advance technologies that have both civilian and defense applications.

Gene Keselman, the executive director of MIx as well as managing director of MIT’s venture studio Proto Ventures and a colonel in the U.S. Air Force Reserve, believes MIT is uniquely positioned to deliver on MIx’s mission.

“It’s not a coincidence MIx is happening at MIT,” says Keselman, adding that supporting national security “is part of MIT’s ethos.”

A history of service

MIx’s work has deep roots at the Institute.

“MIT has worked with the Department of Defense since at least since the 1940s, but really going back to its founding years,” says Karaman, who is also the director of MIT’s Laboratory for Information and Decision Systems (LIDS), a research group with its own long history of working with the government.

“The difference today,” adds Murray, who teaches courses on building deep tech ventures and regional innovation ecosystems and is the vice chair of NATO's Innovation Fund, “is that defense departments and others looking to support the defense, security, and resilience agenda are looking to several innovation ecosystem stakeholders — universities, startup ventures, and venture capitalists — for solutions. Not only from the large prime contractors.  We have learned this lesson from Ukraine, but the same ecosystem logic is at the core of our MIx offer.”

MIx was borne out of the MIT Innovation Initiative in response to interest Keselman saw from researchers and defense officials in expanding MIT’s work with the defense and global security communities. About seven years ago, he hired Katie Person, who left MIT last year to become a battalion commander, to handle all that interest as a program manager with the initiative. MIx activities, like mentoring and educating founders, began shortly after, and MIx officially launched at MIT in 2021.

“It was a good example of the ways in which MIT responds to its students’ interests and external demand,” Keselman says.

One source of early interest was from startup founders who wanted to know how to work with the defense industry and commercialize technology that could have dual commercial and defense applications. That led the team to launch the Dual Use Ventures course, which helps startup founders and other innovators work with defense agencies. The course has since been offered annually during MIT’s Independent Activities Period (IAP) and tailored for NATO’s Defense Innovation Accelerator for the North Atlantic (DIANA).

Personnel from agencies including U.S. Special Operations Command were also interested in working with MIT students, which led the MIx team to develop course 15.362/6.9160 (Engineering Innovation: Global Security Systems), which is taken each spring by students across MIT and Harvard University.

“There are the government organizations that want to be more innovative and work with startups, and there are startups that want to get access to funding from government and have government as a customer,” Keselman says. “We’re kind of the middle layer, facilitating connections, educating, and partnering on research.”

MIx research activities give student and graduate researchers opportunities to work on pressing problems in the real world, and the MIT community has responded eagerly: More than 150 students applied for MIx’s openings in this summer’s Undergraduate Research Opportunities Program.

"We’re helping push the boundaries of what’s possible and explore the frontiers of technology, but do it in a way that is publishable," says MIx Head Research Scientist A.J. Perez ’13, MEng ’14, PhD ’23. “More broadly, we want to unlock as much support for students and researchers at MIT as possible to work on problems that we know matter to defense agencies.”

Early wins

Some of MIx’s most impactful research so far has come in partnership with startups. For example, MIx helped the startup Picogrid secure a small business grant from the U.S. Air Force to build an early wildfire detection system. As part of the grant, MIT students built a computer vision model for Picogrid’s devices that can detect smoke in the sky, proving the technical feasibility of the system and describing a promising new pathway in the field of machine learning.

In another recent project with the MIT alumni-founded startup Nominal, MIT students helped improve and automate post-flight data analysis for the U.S. Air Force’s Test Pilot School.

MIx’s work connecting MIT’s innovators and the wider innovation ecosystem with defense agencies has already begun to bear fruit, and many members of MIx believe early collaborations are a sign of things to come.

“We haven’t even scratched the surface of the potential for MIx,” says Karaman, “This could be the start of something much bigger.”

A large language model (LLM) deployed to make treatment recommendations can be tripped up by nonclinical information in patient messages, like typos, extra white space, missing gender markers, or the use of uncertain, dramatic, and informal language, according to a study by MIT researchers.

They found that making stylistic or grammatical changes to messages increases the likelihood an LLM will recommend that a patient self-manage their reported health condition rather than come in for an appointment, even when that patient should seek medical care.

Their analysis also revealed that these nonclinical variations in text, which mimic how people really communicate, are more likely to change a model’s treatment recommendations for female patients, resulting in a higher percentage of women who were erroneously advised not to seek medical care, according to human doctors.

This work “is strong evidence that models must be audited before use in health care — which is a setting where they are already in use,” says Marzyeh Ghassemi, an associate professor in the MIT Department of Electrical Engineering and Computer Science (EECS), a member of the Institute of Medical Engineering Sciences and the Laboratory for Information and Decision Systems, and senior author of the study.

These findings indicate that LLMs take nonclinical information into account for clinical decision-making in previously unknown ways. It brings to light the need for more rigorous studies of LLMs before they are deployed for high-stakes applications like making treatment recommendations, the researchers say.

“These models are often trained and tested on medical exam questions but then used in tasks that are pretty far from that, like evaluating the severity of a clinical case. There is still so much about LLMs that we don’t know,” adds Abinitha Gourabathina, an EECS graduate student and lead author of the study.

They are joined on the paper, which will be presented at the ACM Conference on Fairness, Accountability, and Transparency, by graduate student Eileen Pan and postdoc Walter Gerych.

Mixed messages

Large language models like OpenAI’s GPT-4 are being used to draft clinical notes and triage patient messages in health care facilities around the globe, in an effort to streamline some tasks to help overburdened clinicians.

A growing body of work has explored the clinical reasoning capabilities of LLMs, especially from a fairness point of view, but few studies have evaluated how nonclinical information affects a model’s judgment.

Interested in how gender impacts LLM reasoning, Gourabathina ran experiments where she swapped the gender cues in patient notes. She was surprised that formatting errors in the prompts, like extra white space, caused meaningful changes in the LLM responses.

To explore this problem, the researchers designed a study in which they altered the model’s input data by swapping or removing gender markers, adding colorful or uncertain language, or inserting extra space and typos into patient messages.

Each perturbation was designed to mimic text that might be written by someone in a vulnerable patient population, based on psychosocial research into how people communicate with clinicians.

For instance, extra spaces and typos simulate the writing of patients with limited English proficiency or those with less technological aptitude, and the addition of uncertain language represents patients with health anxiety.

“The medical datasets these models are trained on are usually cleaned and structured, and not a very realistic reflection of the patient population. We wanted to see how these very realistic changes in text could impact downstream use cases,” Gourabathina says.

They used an LLM to create perturbed copies of thousands of patient notes while ensuring the text changes were minimal and preserved all clinical data, such as medication and previous diagnosis. Then they evaluated four LLMs, including the large, commercial model GPT-4 and a smaller LLM built specifically for medical settings.

They prompted each LLM with three questions based on the patient note: Should the patient manage at home, should the patient come in for a clinic visit, and should a medical resource be allocated to the patient, like a lab test.

The researchers compared the LLM recommendations to real clinical responses.

Inconsistent recommendations

They saw inconsistencies in treatment recommendations and significant disagreement among the LLMs when they were fed perturbed data. Across the board, the LLMs exhibited a 7 to 9 percent increase in self-management suggestions for all nine types of altered patient messages.

This means LLMs were more likely to recommend that patients not seek medical care when messages contained typos or gender-neutral pronouns, for instance. The use of colorful language, like slang or dramatic expressions, had the biggest impact.

They also found that models made about 7 percent more errors for female patients and were more likely to recommend that female patients self-manage at home, even when the researchers removed all gender cues from the clinical context.

Many of the worst results, like patients told to self-manage when they have a serious medical condition, likely wouldn’t be captured by tests that focus on the models’ overall clinical accuracy.

“In research, we tend to look at aggregated statistics, but there are a lot of things that are lost in translation. We need to look at the direction in which these errors are occurring — not recommending visitation when you should is much more harmful than doing the opposite,” Gourabathina says.

The inconsistencies caused by nonclinical language become even more pronounced in conversational settings where an LLM interacts with a patient, which is a common use case for patient-facing chatbots.

But in follow-up work, the researchers found that these same changes in patient messages don’t affect the accuracy of human clinicians.

“In our follow up work under review, we further find that large language models are fragile to changes that human clinicians are not,” Ghassemi says. “This is perhaps unsurprising — LLMs were not designed to prioritize patient medical care. LLMs are flexible and performant enough on average that we might think this is a good use case. But we don’t want to optimize a health care system that only works well for patients in specific groups.”

The researchers want to expand on this work by designing natural language perturbations that capture other vulnerable populations and better mimic real messages. They also want to explore how LLMs infer gender from clinical text.

Launched in February of this year, the MIT Generative AI Impact Consortium (MGAIC), a presidential initiative led by MIT’s Office of Innovation and Strategy and administered by the MIT Stephen A. Schwarzman College of Computing, issued a call for proposals, inviting researchers from across MIT to submit ideas for innovative projects studying high-impact uses of generative AI models.

The call received 180 submissions from nearly 250 faculty members, spanning all of MIT’s five schools and the college. The overwhelming response across the Institute exemplifies the growing interest in AI and follows in the wake of MIT’s Generative AI Week and call for impact papers. Fifty-five proposals were selected for MGAIC’s inaugural seed grants, with several more selected to be funded by the consortium’s founding company members.

Over 30 funding recipients presented their proposals to the greater MIT community at a kickoff event on May 13. Anantha P. Chandrakasan, chief innovation and strategy officer and dean of the School of Engineering who is head of the consortium, welcomed the attendees and thanked the consortium’s founding industry members.

“The amazing response to our call for proposals is an incredible testament to the energy and creativity that MGAIC has sparked at MIT. We are especially grateful to our founding members, whose support and vision helped bring this endeavor to life,” adds Chandrakasan. “One of the things that has been most remarkable about MGAIC is that this is a truly cross-Institute initiative. Deans from all five schools and the college collaborated in shaping and implementing it.”

Vivek F. Farias, the Patrick J. McGovern (1959) Professor at the MIT Sloan School of Management and co-faculty director of the consortium with Tim Kraska, associate professor of electrical engineering and computer science in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), emceed the afternoon of five-minute lightning presentations.

Presentation highlights include:

“AI-Driven Tutors and Open Datasets for Early Literacy Education,” presented by Ola Ozernov-Palchik, a research scientist at the McGovern Institute for Brain Research, proposed a refinement for AI-tutors for pK-7 students to potentially decrease literacy disparities.

“Developing jam_bots: Real-Time Collaborative Agents for Live Human-AI Musical Improvisation,” presented by Anna Huang, assistant professor of music and assistant professor of electrical engineering and computer science, and Joe Paradiso, the Alexander W. Dreyfoos (1954) Professor in Media Arts and Sciences at the MIT Media Lab, aims to enhance human-AI musical collaboration in real-time for live concert improvisation.

“GENIUS: GENerative Intelligence for Urban Sustainability,” presented by Norhan Bayomi, a postdoc at the MIT Environmental Solutions Initiative and a research assistant in the Urban Metabolism Group, which aims to address the critical gap of a standardized approach in evaluating and benchmarking cities’ climate policies.

Georgia Perakis, the John C Head III Dean (Interim) of the MIT Sloan School of Management and professor of operations management, operations research, and statistics, who serves as co-chair of the GenAI Dean’s oversight group with Dan Huttenlocher, dean of the MIT Schwarzman College of Computing, ended the event with closing remarks that emphasized “the readiness and eagerness of our community to lead in this space.”

“This is only the beginning,” he continued. “We are at the front edge of a historic moment — one where MIT has the opportunity, and the responsibility, to shape the future of generative AI with purpose, with excellence, and with care.”

The open space connecting Hockfield Court with Massachusetts Avenue, in the heart of MIT’s campus, is now the L. Rafael Reif Innovation Corridor, in honor of the Institute’s 17th president. At a dedication ceremony Monday, Reif’s colleagues, friends, and family gathered to honor his legacy and unveil a marker for the walkway that was previously known as North Corridor or “the Outfinite.”

“It’s no accident that the space we dedicate today is not a courtyard, but a corridor — a channel for people and ideas to flow freely through the heart of MIT, and to carry us outward, to limits of our aspirations,” said Sally Kornbluth, who succeeded Reif as MIT president in 2023.

“With his signature combination of new-world thinking and old-world charm, and as a grand thinker and doer, Rafael left an indelible mark on MIT,” Kornbluth said. “As a permanent testament to his service and his achievements in service to MIT, the nation, and the world, we now dedicate this space as the L. Rafael Reif Innovation Corridor.”

Reif served as president for more than 10 years, following seven years as provost. He has been at MIT since 1980, when he joined the faculty as an assistant professor of electrical engineering.

“Through all those roles, what stood out most was his humility, his curiosity, and his remarkable ability to speak with clarity and conviction,” said Corporation Chair Mark Gorenberg, who opened the ceremony. “Under his leadership, MIT not only stayed true to its mission, it thrived, expanding its impact and strengthening its global voice.”

Gorenberg introduced Abraham J. Siegel Professor of Management and professor of operations research Cindy Barnhart, who served as chancellor, then provost, during Reif’s term as president. Barnhart, who will be stepping down as provost on July 1, summarized the many highlights from Reif’s presidency, such as the establishment of MIT Schwarzman College of Computing, the revitalization of Kendall Square, and the launch of The Engine, as well as the construction or modernization of many buildings, from the Wright Brothers Wind Tunnel to the new Edward and Joyce Linde Music Building, among other accomplishments.

“Beyond space, Rafael’s bold thinking and passion extends to MIT’s approach to education,” Barnhart continued, describing how Reif championed the building of OpenCourseWare, MITx, and edX. She also noted his support for the health and well-being of the MIT community, through efforts such as addressing student sexual misconduct and forming the MindHandHeart initiative. He also hosted dance parties and socials, joined students in the dining halls for dinner, chatted with faculty and staff over breakfasts and at forums, and more.

“At gatherings over the years, Rafael’s wife, Chris, was there by his side,” Barnhart noted, adding, “I’d like to take this opportunity to acknowledge her and thank her for her welcoming and gracious spirit.”

In summary, “I am grateful to Rafael for his visionary leadership and for his love of MIT and its people,” Barnhart said as she presented Reif with a 3D-printed replica of the Maclaurin buildings (MIT Buildings 3, 4, and 10), which was created through a collaboration between the Glass Lab, Edgerton Center, and Project Manus.

Next, Institute Professor Emeritus John Harbison played an interlude on the piano, and a musical ensemble reprised the “Rhumba for Rafael,” which Harbison composed for Reif’s inauguration in 2012.

When Reif took the podium, he reflected on the location of the corridor and its significance to early chapters in his own career; his first office and lab were in Building 13, overlooking what is now the eponymous walkway.

He also considered the years ahead: “The people who pass through this corridor in the future will surely experience the unparalleled excitement of being young at MIT, with the full expectation of upending the world to improve it,” he said.

Faculty and staff walking through the corridor may experience the “undimmed excitement” of working and studying alongside extraordinary students and colleagues, and feeling the “deep satisfaction of having created infinite memories here throughout a long career.”

“Even if none of them gives me a thought,” Reif continued, “I would like to believe that my spirit will be here, watching them with pride as they continue the never-ending mission of creating a better world.”

Volcanic islands, such as the islands of Hawaii and the Caribbean, are surrounded by coral reefs that encircle an island in a labyrinthine, living ring. A coral reef is punctured at points by reef passes — wide channels that cut through the coral and serve as conduits for ocean water and nutrients to filter in and out. These watery passageways provide circulation throughout a reef, helping to maintain the health of corals by flushing out freshwater and transporting key nutrients.

Now, MIT scientists have found that reef passes are shaped by island rivers. In a study appearing today in the journal Geophysical Research Letters, the team shows that the locations of reef passes along coral reefs line up with where rivers funnel out from an island’s coast.

Their findings provide the first quantitative evidence of rivers forming reef passes.  Scientists and explorers had speculated that this may be the case: Where a river on a volcanic island meets the coast, the freshwater and sediment it carries flows toward the reef, where a strong enough flow can tunnel into the surrounding coral. This idea has been proposed from time to time but never quantitatively tested, until now.

“The results of this study help us to understand how the health of coral reefs depends on the islands they surround,” says study author Taylor Perron, the Cecil and Ida Green Professor of Earth, Atmospheric and Planetary Sciences at MIT.

“A lot of discussion around rivers and their impact on reefs today has been negative because of human impact and the effects of agricultural practices,” adds lead author Megan Gillen, a graduate student in the MIT-WHOI Joint Program in Oceanography. “This study shows the potential long-term benefits rivers can have on reefs, which I hope reshapes the paradigm and highlights the natural state of rivers interacting with reefs.”

The study’s other co-author is Andrew Ashton of the Woods Hole Oceanographic Institution.

Drawing the lines

The new study is based on the team’s analysis of the Society Islands, a chain of islands in the South Pacific Ocean that includes Tahiti and Bora Bora. Gillen, who joined the MIT-WHOI program in 2020, was interested in exploring connections between coral reefs and the islands they surround. With limited options for on-site work during the Covid-19 pandemic, she and Perron looked to see what they could learn through satellite images and maps of island topography. They did a quick search using Google Earth and zeroed in on the Society Islands for their uniquely visible reef and island features.

“The islands in this chain have these iconic, beautiful reefs, and we kept noticing these reef passes that seemed to align with deeply embayed portions of the coastline,” Gillen says. “We started asking ourselves, is there a correlation here?”

Viewed from above, the coral reefs that circle some islands bear what look to be notches, like cracks that run straight through a ring. These breaks in the coral are reef passes — large channels that run tens of meters deep and can be wide enough for some boats to pass through. On first look, Gillen noticed that the most obvious reef passes seemed to line up with flooded river valleys — depressions in the coastline that have been eroded over time by island rivers that flow toward the ocean. She wondered whether and to what extent island rivers might shape reef passes.

“People have examined the flow through reef passes to understand how ocean waves and seawater circulate in and out of lagoons, but there have been no claims of how these passes are formed,” Gillen says. “Reef pass formation has been mentioned infrequently in the literature, and people haven’t explored it in depth.”

Reefs unraveled

To get a detailed view of the topography in and around the Society Islands, the team used data from the NASA Shuttle Radar Topography Mission — two radar antennae that flew aboard the space shuttle in 1999 and measured the topography across 80 percent of the Earth’s surface.

The researchers used the mission’s topographic data in the Society Islands to create a map of every drainage basin along the coast of each island, to get an idea of where major rivers flow or once flowed. They also marked the locations of every reef pass in the surrounding coral reefs. They then essentially “unraveled” each island’s coastline and reef into a straight line, and compared the locations of basins versus reef passes.

“Looking at the unwrapped shorelines, we find a significant correlation in the spatial relationship between these big river basins and where the passes line up,” Gillen says. “So we can say that statistically, the alignment of reef passes and large rivers does not seem random. The big rivers have a role in forming passes.”

As for how rivers shape the coral conduits, the team has two ideas, which they call, respectively, reef incision and reef encroachment. In reef incision, they propose that reef passes can form in times when the sea level is relatively low, such that the reef is exposed above the sea surface and a river can flow directly over the reef. The water and sediment carried by the river can then erode the coral, progressively carving a path through the reef.

When sea level is relatively higher, the team suspects a reef pass can still form, through reef encroachment. Coral reefs naturally live close to the water surface, where there is light and opportunity for photosynthesis. When sea levels rise, corals naturally grow upward and inward toward an island, to try to “catch up” to the water line.

“Reefs migrate toward the islands as sea levels rise, trying to keep pace with changing average sea level,” Gillen says.

However, part of the encroaching reef can end up in old river channels that were previously carved out by large rivers and that are lower than the rest of the island coastline. The corals in these river beds end up deeper than light can extend into the water column, and inevitably drown, leaving a gap in the form of a reef pass.

“We don’t think it’s an either/or situation,” Gillen says. “Reef incision occurs when sea levels fall, and reef encroachment happens when sea levels rise. Both mechanisms, occurring over dozens of cycles of sea-level rise and island evolution, are likely responsible for the formation and maintenance of reef passes over time.”

The team also looked to see whether there were differences in reef passes in older versus younger islands. They observed that younger islands were surrounded by more reef passes that were spaced closer together, versus older islands that had fewer reef passes that were farther apart.

As islands age, they subside, or sink, into the ocean, which reduces the amount of land that funnels rainwater into rivers. Eventually, rivers are too weak to keep the reef passes open, at which point, the ocean likely takes over, and incoming waves could act to close up some passes.

Gillen is exploring ideas for how rivers, or river-like flow, can be engineered to create paths through coral reefs in ways that would promote circulation and benefit reef health.

“Part of me wonders: If you had a more persistent flow, in places where you don’t naturally have rivers interacting with the reef, could that potentially be a way to increase health, by incorporating that river component back into the reef system?” Gillen says. “That’s something we’re thinking about.”

This research was supported, in part, by the WHOI Watson and Von Damm fellowships.

Water makes up around 60 percent of the human body. More than half of this water sloshes around inside the cells that make up organs and tissues. Much of the remaining water flows in the nooks and crannies between cells, much like seawater between grains of sand.

Now, MIT engineers have found that this “intercellular” fluid plays a major role in how tissues respond when squeezed, pressed, or physically deformed. Their findings could help scientists understand how cells, tissues, and organs physically adapt to conditions such as aging, cancer, diabetes, and certain neuromuscular diseases.

In a paper appearing today in Nature Physics, the researchers show that when a tissue is pressed or squeezed, it is more compliant and relaxes more quickly when the fluid between its cells flows easily. When the cells are packed together and there is less room for intercellular flow, the tissue as a whole is stiffer and resists being pressed or squeezed.

The findings challenge conventional wisdom, which has assumed that a tissue’s compliance depends mainly on what’s inside, rather than around, a cell. Now that the researchers have shown that intercellular flow determines how tissues will adapt to physical forces, the results can be applied to understand a wide range of physiological conditions, including how muscles withstand exercise and recover from injury, and how a tissue’s physical adaptability may affect the progression of aging, cancer, and other medical conditions.

The team envisions the results could also inform the design of artificial tissues and organs. For instance, in engineering artificial tissue, scientists might optimize intercellular flow within the tissue to improve its function or resilience. The researchers suspect that intercellular flow could also be a route for delivering nutrients or therapies, either to heal a tissue or eradicate a tumor.

“People know there is a lot of fluid between cells in tissues, but how important that is, in particular in tissue deformation, is completely ignored,” says Ming Guo, associate professor of mechanical engineering at MIT. “Now we really show we can observe this flow. And as the tissue deforms, flow between cells dominates the behavior. So, let’s pay attention to this when we study diseases and engineer tissues.”

Guo is a co-author of the new study, which includes lead author and MIT postdoc Fan Liu PhD ’24, along with Bo Gao and Hui Li of Beijing Normal University, and Liran Lei and Shuainan Liu of Peking Union Medical College.

Pressed and squeezed

The tissues and organs in our body are constantly undergoing physical deformations, from the large stretch and strain of muscles during motion to the small and steady contractions of the heart. In some cases, how easily tissues adapt to deformation can relate to how quickly a person can recover from, for instance, an allergic reaction, a sports injury, or a brain stroke. However, exactly what sets a tissue’s response to deformation is largely unknown.

Guo and his group at MIT looked into the mechanics of tissue deformation, and the role of intercellular flow in particular, following a study they published in 2020. In that study, they focused on tumors and observed the way in which fluid can flow from the center of a tumor out to its edges, through the cracks and crevices between individual tumor cells. They found that when a tumor was squeezed or pressed, the intercellular flow increased, acting as a conveyor belt to transport fluid from the center to the edges. Intercellular flow, they found, could fuel tumor invasion into surrounding regions.

In their new study, the team looked to see what role this intercellular flow might play in other, noncancerous tissues.

“Whether you allow the fluid to flow between cells or not seems to have a major impact,” Guo says. “So we decided to look beyond tumors to see how this flow influences how other tissues respond to deformation.”

A fluid pancake

Guo, Liu, and their colleagues studied the intercellular flow in a variety of biological tissues, including cells derived from pancreatic tissue. They carried out experiments in which they first cultured small clusters of tissue, each measuring less than a quarter of a millimeter wide and numbering tens of thousands of individual cells. They placed each tissue cluster in a custom-designed testing platform that the team built specifically for the study.

“These microtissue samples are in this sweet zone where they are too large to see with atomic force microscopy techniques and too small for bulkier devices,” Guo says. “So, we decided to build a device.”

The researchers adapted a high-precision microbalance that measures minute changes in weight. They combined this with a step motor that is designed to press down on a sample with nanometer precision. The team placed tissue clusters one at a time on the balance and recorded each cluster’s changing weight as it relaxed from a sphere into the shape of a pancake in response to the compression. The team also took videos of the clusters as they were squeezed.

For each type of tissue, the team made clusters of varying sizes. They reasoned that if the tissue’s response is ruled by the flow between cells, then the bigger a tissue, the longer it should take for water to seep through, and therefore, the longer it should take the tissue to relax. It should take the same amount of time, regardless of size, if a tissue’s response is determined by the structure of the tissue rather than fluid.

Over multiple experiments with a variety of tissue types and sizes, the team observed a similar trend: The bigger the cluster, the longer it took to relax, indicating that intercellular flow dominates a tissue’s response to deformation.

“We show that this intercellular flow is a crucial component to be considered in the fundamental understanding of tissue mechanics and also applications in engineering living systems,” Liu says.

Going forward, the team plans to look into how intercellular flow influences brain function, particularly in disorders such as Alzheimer’s disease.

“Intercellular or interstitial flow can help you remove waste and deliver nutrients to the brain,” Liu adds. “Enhancing this flow in some cases might be a good thing.”

“As this work shows, as we apply pressure to a tissue, fluid will flow,” Guo says. “In the future, we can think of designing ways to massage a tissue to allow fluid to transport nutrients between cells.”

More than half of the nation’s 623,218 bridges are experiencing significant deterioration. Through an in-field case study conducted in western Massachusetts, a team led by the University of Massachusetts at Amherst in collaboration with researchers from the MIT Department of Mechanical Engineering (MechE) has just successfully demonstrated that 3D printing may provide a cost-effective, minimally disruptive solution.

“Anytime you drive, you go under or over a corroded bridge,” says Simos Gerasimidis, associate professor of civil and environmental engineering at UMass Amherst and former visiting professor in the Department of Civil and Environmental Engineering at MIT, in a press release. “They are everywhere. It’s impossible to avoid, and their condition often shows significant deterioration. We know the numbers.”

The numbers, according to the American Society of Civil Engineers’ 2025 Report Card for America’s Infrastructure, are staggering: Across the United States, 49.1 percent of the nation’s 623,218 bridges are in “fair” condition and 6.8 percent are in “poor” condition. The projected cost to restore all of these failing bridges exceeds $191 billion.

A proof-of-concept repair took place last month on a small, corroded section of a bridge in Great Barrington, Massachusetts. The technique, called cold spray, can extend the life of beams, reinforcing them with newly deposited steel. The process accelerates particles of powdered steel in heated, compressed gas, and then a technician uses an applicator to spray the steel onto the beam. Repeated sprays create multiple layers, restoring thickness and other structural properties.

This method has proven to be an effective solution for other large structures like submarines, airplanes, and ships, but bridges present a problem on a greater scale. Unlike movable vessels, stationary bridges cannot be brought to the 3D printer — the printer must be brought on-site — and, to lessen systemic impacts, repairs must also be made with minimal disruptions to traffic, which the new approach allows.

“Now that we’ve completed this proof-of-concept repair, we see a clear path to a solution that is much faster, less costly, easier, and less invasive,” says Gerasimidis. “To our knowledge, this is a first. Of course, there is some R&D that needs to be developed, but this is a huge milestone to that.”

“This is a tremendous collaboration where cutting-edge technology is brought to address a critical need for infrastructure in the commonwealth and across the United States,” says John Hart, Class of 1922 Professor and head of the Department of MechE at MIT. Hart and Haden Quinlan, senior program manager in the Center for Advanced Production Technologies at MIT, are leading MIT’s efforts in in the project. Hart is also faculty co-lead of the recently announced MIT Initiative for New Manufacturing.

“Integrating digital systems with advanced physical processing is the future of infrastructure,” says Quinlan. “We’re excited to have moved this technology beyond the lab and into the field, and grateful to our collaborators in making this work possible.”

UMass says the Massachusetts Department of Transportation (MassDOT) has been a valued research partner, helping to identify the problem and providing essential support for the development and demonstration of the technology. Technical guidance and funding support were provided by the MassDOT Highway Division and the Research and Technology Transfer Program.

Equipment for this project was supported through the Massachusetts Manufacturing Innovation Initiative, a statewide program led by the Massachusetts Technology Collaborative (MassTech)’s Center for Advanced Manufacturing that helps bridge the gap between innovation and commercialization in hard tech manufacturing.

“It’s a very Massachusetts success story,” Gerasimidis says. “It involves MassDOT being open-minded to new ideas. It involves UMass and MIT putting [together] the brains to do it. It involves MassTech to bring manufacturing back to Massachusetts. So, I think it’s a win-win for everyone involved here.”

The bridge in Great Barrington is scheduled for demolition in a few years. After demolition occurs, the recently-sprayed beams will be taken back to UMass for testing and measurement to study how well the deposited steel powder adhered to the structure in the field compared to in a controlled lab setting, if it corroded further after it was sprayed, and determine its mechanical properties.

This demonstration builds on several years of research by the UMass and MIT teams, including development of a “digital thread” approach to scan corroded beam surfaces and determine material deposition profiles, alongside laboratory studies of cold spray and other additive manufacturing approaches that are suited to field deployment.

Altogether, this work is a collaborative effort among UMass Amherst, MIT MechE, MassDOT, the Massachusetts Technology Collaborative (MassTech), the U.S. Department of Transportation, and the Federal Highway Administration. Research reports are available on the MassDOT website.  

When the Earth froze over, where did life shelter? MIT scientists say one refuge may have been pools of melted ice that dotted the planet’s icy surface.

In a study appearing today in Nature Communications, the researchers report that 635 million to 720 million years ago, during periods known as “Snowball Earth,” when much of the planet was covered in ice, some of our ancient cellular ancestors could have waited things out in meltwater ponds.

The scientists found that eukaryotes — complex cellular lifeforms that eventually evolved into the diverse multicellular life we see today — could have survived the global freeze by living in shallow pools of water. These small, watery oases may have persisted atop relatively shallow ice sheets present in equatorial regions. There, the ice surface could accumulate dark-colored dust and debris from below, which enhanced its ability to melt into pools. At temperatures hovering around 0 degrees Celsius, the resulting meltwater ponds could have served as habitable environments for certain forms of early complex life.

The team drew its conclusions based on an analysis of modern-day meltwater ponds. Today in Antarctica, small pools of melted ice can be found along the margins of ice sheets. The conditions along these polar ice sheets are similar to what likely existed along ice sheets near the equator during Snowball Earth.

The researchers analyzed samples from a variety of meltwater ponds located on the McMurdo Ice Shelf in an area that was first described by members of Robert Falcon Scott's 1903 expedition as “dirty ice.” The MIT researchers discovered clear signatures of eukaryotic life in every pond. The communities of eukaryotes varied from pond to pond, revealing a surprising diversity of life across the setting. The team also found that salinity plays a key role in the kind of life a pond can host: Ponds that were more brackish or salty had more similar eukaryotic communities, which differed from those in ponds with fresher waters.

“We’ve shown that meltwater ponds are valid candidates for where early eukaryotes could have sheltered during these planet-wide glaciation events,” says lead author Fatima Husain, a graduate student in MIT’s Department of Earth, Atmospheric and Planetary Sciences (EAPS). “This shows us that diversity is present and possible in these sorts of settings. It’s really a story of life’s resilience.”

The study’s MIT co-authors include Schlumberger Professor of Geobiology Roger Summons and former postdoc Thomas Evans, along with Jasmin Millar of Cardiff University, Anne Jungblut at the Natural History Museum in London, and Ian Hawes of the University of Waikato in New Zealand.

Polar plunge

“Snowball Earth” is the colloquial term for periods of time in Earth history during which the planet iced over. It is often used as a reference to the two consecutive, multi-million-year glaciation events which took place during the Cryogenian Period, which geologists refer to as the time between 635 and 720 million years ago. Whether the Earth was more of a hardened snowball or a softer “slushball” is still up for debate. But scientists are certain of one thing: Most of the planet was plunged into a deep freeze, with average global temperatures of minus 50 degrees Celsius. The question has been: How and where did life survive?

“We’re interested in understanding the foundations of complex life on Earth. We see evidence for eukaryotes before and after the Cryogenian in the fossil record, but we largely lack direct evidence of where they may have lived during,” Husain says. “The great part of this mystery is, we know life survived. We’re just trying to understand how and where.”

There are a number of ideas for where organisms could have sheltered during Snowball Earth, including in certain patches of the open ocean (if such environments existed), in and around deep-sea hydrothermal vents, and under ice sheets. In considering meltwater ponds, Husain and her colleagues pursued the hypothesis that surface ice meltwaters may also have been capable of supporting early eukaryotic life at the time.

“There are many hypotheses for where life could have survived and sheltered during the Cryogenian, but we don’t have excellent analogs for all of them,” Husain notes. “Above-ice meltwater ponds occur on Earth today and are accessible, giving us the opportunity to really focus in on the eukaryotes which live in these environments.”

Small pond, big life

For their new study, the researchers analyzed samples taken from meltwater ponds in Antarctica. In 2018, Summons and colleagues from New Zealand traveled to a region of the McMurdo Ice Shelf in East Antarctica, known to host small ponds of melted ice, each just a few feet deep and a few meters wide. There, water freezes all the way to the seafloor, in the process trapping dark-colored sediments and marine organisms. Wind-driven loss of ice from the surface creates a sort of conveyer belt that brings this trapped debris to the surface over time, where it absorbs the sun’s warmth, causing ice to melt, while surrounding debris-free ice reflects incoming sunlight, resulting in the formation of shallow meltwater ponds.

The bottom of each pond is lined with mats of microbes that have built up over years to form layers of sticky cellular communities.

“These mats can be a few centimeters thick, colorful, and they can be very clearly layered,” Husain says.

These microbial mats are made up of cyanobacteria, prokaryotic, single-celled photosynthetic organisms that lack a cell nucleus or other organelles. While these ancient microbes are known to survive within some of the the harshest environments on Earth including meltwater ponds, the researchers wanted to know whether eukaryotes — complex organisms that evolved a cell nucleus and other membrane bound organelles — could also weather similarly challenging circumstances. Answering this question would take more than a microscope, as the defining characteristics of the microscopic eukaryotes present among the microbial mats are too subtle to distinguish by eye.

To characterize the eukaryotes, the team analyzed the mats for specific lipids they make called sterols, as well as genetic components called ribosomal ribonucleic acid (rRNA), both of which can be used to identify organisms with varying degrees of specificity. These two independent sets of analyses provided complementary fingerprints for certain eukaryotic groups. As part of the team’s lipid research, they found many sterols and rRNA genes closely associated with specific types of algae, protists, and microscopic animals among the microbial mats. The researchers were able to assess the types and relative abundance of lipids and rRNA genes from pond to pond, and found the ponds hosted a surprising diversity of eukaryotic life.

“No two ponds were alike,” Husain says. “There are repeating casts of characters, but they’re present in different abundances. And we found diverse assemblages of eukaryotes from all the major groups in all the ponds studied. These eukaryotes are the descendants of the eukaryotes that survived the Snowball Earth. This really highlights that meltwater ponds during Snowball Earth could have served as above-ice oases that nurtured the eukaryotic life that enabled the diversification and proliferation of complex life — including us — later on.”

This research was supported, in part, by the NASA Exobiology Program, the Simons Collaboration on the Origins of Life, and a MISTI grant from MIT-New Zealand.

MIT has again been named the world’s top university by the QS World University Rankings, which were announced today. This is the 14th year in a row MIT has received this distinction.

The full 2026 edition of the rankings — published by Quacquarelli Symonds, an organization specializing in education and study abroad — can be found at TopUniversities.com. The QS rankings are based on factors including academic reputation, employer reputation, citations per faculty, student-to-faculty ratio, proportion of international faculty, and proportion of international students.

MIT was also ranked the world’s top university in 11 of the subject areas ranked by QS, as announced in March of this year.

The Institute received a No. 1 ranking in the following QS subject areas: Chemical Engineering; Civil and Structural Engineering; Computer Science and Information Systems; Data Science and Artificial Intelligence; Electrical and Electronic Engineering; Linguistics; Materials Science; Mechanical, Aeronautical, and Manufacturing Engineering; Mathematics; Physics and Astronomy; and Statistics and Operational Research.

MIT also placed second in seven subject areas: Accounting and Finance; Architecture/Built Environment; Biological Sciences; Business and Management Studies; Chemistry; Earth and Marine Sciences; and Economics and Econometrics.

Social security numbers stolen. Public transport halted. Hospital systems frozen until ransoms are paid. These are some of the damaging consequences of unsecure memory in computer systems.

Over the past decade, public awareness of such cyberattacks has intensified, as their impacts have harmed individuals, corporations, and governments. Today, this awareness is coinciding with technologies that are finally mature enough to eliminate vulnerabilities in memory safety.  

"We are at a tipping point — now is the right time to move to memory-safe systems," says Hamed Okhravi, a cybersecurity expert in MIT Lincoln Laboratory’s Secure Resilient Systems and Technology Group.

In an op-ed earlier this year in Communications of the ACM, Okhravi joined 20 other luminaries in the field of computer security to lay out a plan for achieving universal memory safety. They argue for a standardized framework as an essential next step to adopting memory-safety technologies throughout all forms of computer systems, from fighter jets to cell phones.

Memory-safety vulnerabilities occur when a program performs unintended or erroneous operations in memory. Such operations are prevalent, accounting for an estimated 70 percent of software vulnerabilities. If attackers gain access to memory, they can potentially steal sensitive information, alter program execution, or even take control of the computer system.

These vulnerabilities exist largely because common software programming languages, such as C or C++, are inherently memory-insecure. A simple error by a software engineer, perhaps one line in a system’s multimillion lines of code, could be enough for an attacker to exploit. In recent years, new memory-safe languages, such as Rust, have been developed. But rewriting legacy systems in new, memory-safe languages can be costly and complicated.

Okhravi focuses on the national security implications of memory-safety vulnerabilities. For the U.S. Department of Defense (DoD), whose systems comprise billions of lines of legacy C or C++ code, memory safety has long been a known problem. The National Security Agency (NSA) and the federal government have recently urged technology developers to eliminate memory-safety vulnerabilities from their products. Security concerns extend beyond military systems to widespread consumer products.

"Cell phones, for example, are not immediately important for defense or war-fighting, but if we have 200 million vulnerable cell phones in the nation, that’s a serious matter of national security," Okhravi says.

Memory-safe technology

In recent years, several technologies have emerged to help patch memory vulnerabilities in legacy systems. As the guest editor for a special issue of IEEE Security and Privacy, Okhravi solicited articles from top contributors in the field to highlight these technologies and the ways they can build on one another.  

Some of these memory-safety technologies have been developed at Lincoln Laboratory, with sponsorship from DoD agencies. These technologies include TRACER and TASR, which are software products for Windows and Linux systems, respectively, that reshuffle the location of code in memory each time a program accesses it, making it very difficult for attackers to find exploits. These moving-target solutions have since been licensed by cybersecurity and cloud services companies.

"These technologies are quick wins, enabling us to make a lot of immediate impact without having to rebuild the whole system. But they are only a partial solution, a way of securing legacy systems while we are transitioning to safer languages," Okhravi says.

Innovative work is underway to make that transition easier. For example, the TRACTOR program at the U.S. Defense Advanced Research Projects Agency is developing artificial intelligence tools to automatically translate legacy C code to Rust. Lincoln Laboratory researchers will test and evaluate the translator for use in DoD systems.

Okhravi and his coauthors acknowledged in their op-ed that the timeline for full adoption of memory-safe systems is long — likely decades. It will require the deployment of a combination of new hardware, software, and techniques, each with their own adoption paths, costs, and disruptions. Organizations should prioritize mission-critical systems first.

"For example, the most important components in a fighter jet, such as the flight-control algorithm or the munition-handling logic, would be made memory-safe, say, within five years," Okhravi says. Subsystems less important to critical functions would have a longer time frame.

Use of memory-safe programming languages at Lincoln Laboratory

As Lincoln Laboratory continues its leadership in advancing memory-safety technologies, the Secure Resilient Systems and Technology Group has prioritized adopting memory-safe programming languages. "We’ve been investing in the group-wide use of Rust for the past six years as part of our broader strategy to prototype cyber-hardened mission systems and high-assurance cryptographic implementations for the DoD and intelligence community," says Roger Khazan, who leads the group. "Memory safety is fundamental to trustworthiness in these systems."

Rust’s strong guarantees around memory safety, along with its speed and ability to catch bugs early during development, make it especially well-suited for building secure and reliable systems. The laboratory has been using Rust to prototype and transition secure components for embedded, distributed, and cryptographic systems where resilience, performance, and correctness are mission-critical.

These efforts support both immediate U.S. government needs and a longer-term transformation of the national security software ecosystem. "They reflect Lincoln Laboratory’s broader mission of advancing technology in service to national security, grounded in technical excellence, innovation, and trust," Khazan adds.

A technology-agnostic framework

As new computer systems are designed, developers need a framework of memory-safety standards guiding them. Today, attempts to request memory safety in new systems are hampered by the lack of a clear set of definitions and practice.

Okhravi emphasizes that this standardized framework should be technology-agnostic and provide specific timelines with sets of requirements for different types of systems.

"In the acquisition process for the DoD, and even the commercial sector, when we are mandating memory safety, it shouldn’t be tied to a specific technology. It should be generic enough that different types of systems can apply different technologies to get there," he says.

Filling this gap not only requires building industrial consensus on technical approaches, but also collaborating with government and academia to bring this effort to fruition.

The need for collaboration was an impetus for the op-ed, and Okhravi says that the consortium of experts will push for standardization from their positions across industry, government, and academia. Contributors to the paper represent a wide range of institutes, from the University of Cambridge and SRI International to Microsoft and Google. Together, they are building momentum to finally root out memory vulnerabilities and the costly damages associated with them.

"We are seeing this cost-risk trade-off mindset shifting, partly because of the maturation of technology and partly because of such consequential incidents,” Okhravi says. "We hear all the time that such-and-such breach cost billions of dollars. Meanwhile, making the system secure might have cost 10 million dollars. Wouldn’t we have been better off making that effort?"

The MIT Press announces the acquisition of textbook publisher University Science Books from AIP Publishing, a subsidiary of the American Institute of Physics (AIP).

University Science Books was founded in 1978 to publish intermediate- and advanced-level science and reference books by respected authors, published with the highest design and production standards, and priced as affordably as possible. Over the years, USB’s authors have acquired international followings, and its textbooks in chemistry, physics, and astronomy have been recognized as the gold standard in their respective disciplines. USB was acquired by AIP Publishing in 2021.

Bestsellers include John Taylor’s “Classical Mechanics,” the No. 1 adopted text for undergrad mechanics courses in the United States and Canada, and his “Introduction to Error Analysis;” and Don McQuarrie’s “Physical Chemistry: A Molecular Approach” (commonly known as “Big Red”), the second-most adopted physical chemistry textbook in the U.S.

“We are so pleased to have found a new home for USB’s prestigious list of textbooks in the sciences,” says Alix Vance, CEO of AIP Publishing. “With its strong STEM focus, academic rigor, and high production standards, the MIT Press is the perfect partner to continue the publishing legacy of University Science Books.” 

“This acquisition is both a brand and content fit for the MIT Press,” says Amy Brand, director and publisher of the MIT Press. “USB’s respected science list will complement our long-established publishing history of publishing foundational texts in computer science, finance, and economics.”

The MIT Press will take over the USB list as of July 1, with inventory transferring to Penguin Random House Publishing Services, the MIT Press’ sales and distribution partner.

For details regarding University Science Books titles, inventory, and how to order, please contact the MIT Press. 

Established in 1962, The MIT Press is one of the largest and most distinguished university presses in the world and a leading publisher of books and journals at the intersection of science, technology, art, social science, and design.

AIP Publishing is a wholly owned not-for-profit subsidiary of the AIP and supports the charitable, scientific, and educational purposes of AIP through scholarly publishing activities on its behalf and on behalf of our publishing partners.

Researchers at MIT and the Scripps Research Institute have shown that they can generate a strong immune response to HIV with just one vaccine dose, by adding two powerful adjuvants — materials that help stimulate the immune system.

In a study of mice, the researchers showed that this approach produced a much wider diversity of antibodies against an HIV antigen, compared to the vaccine given on its own or with just one of the adjuvants. The dual-adjuvant vaccine accumulated in the lymph nodes and remained there for up to a month, allowing the immune system to build up a much greater number of antibodies against the HIV protein.

This strategy could lead to the development of vaccines that only need to be given once, for infectious diseases including HIV or SARS-CoV-2, the researchers say.

“This approach is compatible with many protein-based vaccines, so it offers the opportunity to engineer new formulations for these types of vaccines across a wide range of different diseases, such as influenza, SARS-CoV-2, or other pandemic outbreaks,” says J. Christopher Love, the Raymond A. and Helen E. St. Laurent Professor of Chemical Engineering at MIT, and a member of the Koch Institute for Integrative Cancer Research and the Ragon Institute of MGH, MIT, and Harvard.

Love and Darrell Irvine, a professor of immunology and microbiology at the Scripps Research Institute, are the senior authors of the study, which appears today in Science Translational Medicine. Kristen Rodrigues PhD ’23 and Yiming Zhang PhD ’25 are the lead authors of the paper.

More powerful vaccines

Most vaccines are delivered along with adjuvants, which help to stimulate a stronger immune response to the antigen. One adjuvant commonly used with protein-based vaccines, including those for hepatitis A and B, is aluminum hydroxide, also known as alum. This adjuvant works by activating the innate immune response, helping the body to form a stronger memory of the vaccine antigen.

Several years ago, Irvine developed another adjuvant based on saponin, an FDA-approved adjuvant derived from the bark of the Chilean soapbark tree. His work showed that nanoparticles containing both saponin and a molecule called MPLA, which promotes inflammation, worked better than saponin on its own. That nanoparticle, known as SMNP, is now being used as an adjuvant for an HIV vaccine that is currently in clinical trials.

Irvine and Love then tried combining alum and SMNP and showed that vaccines containing both of those adjuvants could generate even more powerful immune responses against either HIV or SARS-CoV-2.

In the new paper, the researchers wanted to explore why these two adjuvants work so well together to boost the immune response, specifically the B cell response. B cells produce antibodies that can circulate in the bloodstream and recognize a pathogen if the body is exposed to it again.

For this study, the researchers used an HIV protein called MD39 as their vaccine antigen, and anchored dozens of these proteins to each alum particle, along with SMNP.

After vaccinating mice with these particles, the researchers found that the vaccine accumulated in the lymph nodes — structures where B cells encounter antigens and undergo rapid mutations that generate antibodies with high affinity for a particular antigen. This process takes place within clusters of cells known as germinal centers.

The researchers showed that SMNP and alum helped the HIV antigen to penetrate through the protective layer of cells surrounding the lymph nodes without being broken down into fragments. The adjuvants also helped the antigens to remain intact in the lymph nodes for up to 28 days.

“As a result, the B cells that are cycling in the lymph nodes are constantly being exposed to the antigen over that time period, and they get the chance to refine their solution to the antigen,” Love says.

This approach may mimic what occurs during a natural infection, when antigens can remain in the lymph nodes for weeks, giving the body time to build up an immune response.

Antibody diversity

Single-cell RNA sequencing of B cells from the vaccinated mice revealed that the vaccine containing both adjuvants generated a much more diverse repertoire of B cells and antibodies. Mice that received the dual-adjuvant vaccine produced two to three times more unique B cells than mice that received just one of the adjuvants.

That increase in B cell number and diversity boosts the chances that the vaccine could generate broadly neutralizing antibodies — antibodies that can recognize a variety of strains of a given virus, such as HIV.

“When you think about the immune system sampling all of the possible solutions, the more chances we give it to identify an effective solution, the better,” Love says. “Generating broadly neutralizing antibodies is something that likely requires both the kind of approach that we showed here, to get that strong and diversified response, as well as antigen design to get the right part of the immunogen shown.”

Using these two adjuvants together could also contribute to the development of more potent vaccines against other infectious diseases, with just a single dose.

“What’s potentially powerful about this approach is that you can achieve long-term exposures based on a combination of adjuvants that are already reasonably well-understood, so it doesn’t require a different technology. It’s just combining features of these adjuvants to enable low-dose or potentially even single-dose treatments,” Love says.

The research was funded by the National Institutes of Health; the Koch Institute Support (core) Grant from the National Cancer Institute; the Ragon Institute of MGH, MIT, and Harvard; and the Howard Hughes Medical Institute.

The advanced semiconductor material gallium nitride will likely be key for the next generation of high-speed communication systems and the power electronics needed for state-of-the-art data centers.

Unfortunately, the high cost of gallium nitride (GaN) and the specialization required to incorporate this semiconductor material into conventional electronics have limited its use in commercial applications.

Now, researchers from MIT and elsewhere have developed a new fabrication process that integrates high-performance GaN transistors onto standard silicon CMOS chips in a way that is low-cost and scalable, and compatible with existing semiconductor foundries.

Their method involves building many tiny transistors on the surface of a GaN chip, cutting out each individual transistor, and then bonding just the necessary number of transistors onto a silicon chip using a low-temperature process that preserves the functionality of both materials.

The cost remains minimal since only a tiny amount of GaN material is added to the chip, but the resulting device can receive a significant performance boost from compact, high-speed transistors. In addition, by separating the GaN circuit into discrete transistors that can be spread over the silicon chip, the new technology is able to reduce the temperature of the overall system.

The researchers used this process to fabricate a power amplifier, an essential component in mobile phones, that achieves higher signal strength and efficiencies than devices with silicon transistors. In a smartphone, this could improve call quality, boost wireless bandwidth, enhance connectivity, and extend battery life.

Because their method fits into standard procedures, it could improve electronics that exist today as well as future technologies. Down the road, the new integration scheme could even enable quantum applications, as GaN performs better than silicon at the cryogenic temperatures essential for many types of quantum computing.

“If we can bring the cost down, improve the scalability, and, at the same time, enhance the performance of the electronic device, it is a no-brainer that we should adopt this technology. We’ve combined the best of what exists in silicon with the best possible gallium nitride electronics. These hybrid chips can revolutionize many commercial markets,” says Pradyot Yadav, an MIT graduate student and lead author of a paper on this method.

He is joined on the paper by fellow MIT graduate students Jinchen Wang and Patrick Darmawi-Iskandar; MIT postdoc John Niroula; senior authors Ulrich L. Rohde, a visiting scientist at the Microsystems Technology Laboratories (MTL), and Ruonan Han, an associate professor in the Department of Electrical Engineering and Computer Science (EECS) and member of MTL; and Tomás Palacios, the Clarence J. LeBel Professor of EECS, and director of MTL; as well as collaborators at Georgia Tech and the Air Force Research Laboratory. The research was recently presented at the IEEE Radio Frequency Integrated Circuits Symposium.

Swapping transistors

Gallium nitride is the second most widely used semiconductor in the world, just after silicon, and its unique properties make it ideal for applications such as lighting, radar systems and power electronics.

The material has been around for decades and, to get access to its maximum performance, it is important for chips made of GaN to be connected to digital chips made of silicon, also called CMOS chips. To enable this, some integration methods bond GaN transistors onto a CMOS chip by soldering the connections, but this limits how small the GaN transistors can be. The tinier the transistors, the higher the frequency at which they can work.

Other methods integrate an entire gallium nitride wafer on top of a silicon wafer, but using so much material is extremely costly, especially since the GaN is only needed in a few tiny transistors. The rest of the material in the GaN wafer is wasted.

“We wanted to combine the functionality of GaN with the power of digital chips made of silicon, but without having to compromise on either cost of bandwidth. We achieved that by adding super-tiny discrete gallium nitride transistors right on top of the silicon chip,” Yadav explains.

The new chips are the result of a multistep process.

First, a tightly packed collection of miniscule transistors is fabricated across the entire surface of a GaN wafer. Using very fine laser technology, they cut each one down to just the size of the transistor, which is 240 by 410 microns, forming what they call a dielet. (A micron is one millionth of a meter.)

Each transistor is fabricated with tiny copper pillars on top, which they use to bond directly to the copper pillars on the surface of a standard silicon CMOS chip. Copper to copper bonding can be done at temperatures below 400 degrees Celsius, which is low enough to avoid damaging either material.

Current GaN integration techniques require bonds that utilize gold, an expensive material that needs much higher temperatures and stronger bonding forces than copper. Since gold can contaminate the tools used in most semiconductor foundries, it typically requires specialized facilities.

“We wanted a process that was low-cost, low-temperature, and low-force, and copper wins on all of those related to gold. At the same time, it has better conductivity,” Yadav says.

A new tool

To enable the integration process, they created a specialized new tool that can carefully integrate the extremely tiny GaN transistor with the silicon chips. The tool uses a vacuum to hold the dielet as it moves on top of a silicon chip, zeroing in on the copper bonding interface with nanometer precision.

They used advanced microscopy to monitor the interface, and then when the dielet is in the right position, they apply heat and pressure to bond the GaN transistor to the chip.

“For each step in the process, I had to find a new collaborator who knew how to do the technique that I needed, learn from them, and then integrate that into my platform. It was two years of constant learning,” Yadav says.

Once the researchers had perfected the fabrication process, they demonstrated it by developing power amplifiers, which are radio frequency circuits that boost wireless signals.

Their devices achieved higher bandwidth and better gain than devices made with traditional silicon transistors. Each compact chip has an area of less than half a square millimeter.

In addition, because the silicon chip they used in their demonstration is based on Intel 16 22nm FinFET state-of-the-art metallization and passive options, they were able to incorporate components often used in silicon circuits, such as neutralization capacitors. This significantly improved the gain of the amplifier, bringing it one step closer to enabling the next generation of wireless technologies.

“To address the slowdown of Moore’s Law in transistor scaling, heterogeneous integration has emerged as a promising solution for continued system scaling, reduced form factor, improved power efficiency, and cost optimization. Particularly in wireless technology, the tight integration of compound semiconductors with silicon-based wafers is critical to realizing unified systems of front-end integrated circuits, baseband processors, accelerators, and memory for next-generation antennas-to-AI platforms. This work makes a significant advancement by demonstrating 3D integration of multiple GaN chips with silicon CMOS and pushes the boundaries of current technological capabilities,” says Atom Watanabe, a research scientist at IBM who was not involved with this paper.

This work is supported, in part, by the U.S. Department of Defense through the National Defense Science and Engineering Graduate (NDSEG) Fellowship Program and CHIMES, one of the seven centers in JUMP 2.0, a Semiconductor Research Corporation Program by the Department of Defense and the Defense Advanced Research Projects Agency (DARPA). Fabrication was carried out using facilities at MIT.Nano, the Air Force Research Laboratory, and Georgia Tech.

MIT Morningside Academy for Design (MAD) Fellow Caitlin Morris is an architect, artist, researcher, and educator who has studied psychology and used online learning tools to teach herself coding and other skills. She’s a soft-spoken observer, with a keen interest in how people use space and respond to their environments. Combining her observational skills with active community engagement, she works at the intersection of technology, education, and human connection to improve digital learning platforms.

Morris grew up in rural upstate New York in a family of makers. She learned to sew, cook, and build things with wood at a young age. One of her earlier memories is of a small handsaw she made — with the help of her father, a professional carpenter. It had wooden handles on both sides to make sawing easier for her.

Later, when she needed to learn something, she’d turn to project-based communities, rather than books. She taught herself to code late at night, taking advantage of community-oriented platforms where people answer questions and post sketches, allowing her to see the code behind the objects people made.

“For me, that was this huge, wake-up moment of feeling like there was a path to expression that was not a traditional computer-science classroom,” she says. “I think that’s partly why I feel so passionate about what I’m doing now. That was the big transformation: having that community available in this really personal, project-based way.”

Subsequently, Morris has become involved in community-based learning in diverse ways: She’s a co-organizer of the MIT Media Lab’s Festival of Learning; she leads creative coding community meetups; and she’s been active in the open-source software community development.

“My years of organizing learning and making communities — both in person and online — have shown me firsthand how powerful social interaction can be for motivation and curiosity,” Morris said. “My research is really about identifying which elements of that social magic are most essential, so we can design digital environments that better support those dynamics.”

Even in her artwork, Morris sometimes works with a collective. She’s contributed to the creation of about 10 large art installations that combine movement, sound, imagery, lighting, and other technologies to immerse the visitor in an experience evoking some aspect of nature, such as flowing water, birds in flight, or crowd kinetics. These marvelous installations are commanding and calming at the same time, possibly because they focus the mind, eye, and sometimes the ear.

She did much of this work with New York-based Hypersonic, a company of artists and technologists specializing in large kinetic installations in public spaces. Before that, she earned a BS in psychology and a BS in architectural building sciences from Rensselaer Polytechnic Institute, then an MFA in design and technology from the Parsons School of Design at The New School.

During, in between, after, and sometimes concurrently, she taught design, coding, and other technologies at the high school, undergraduate, and graduate-student levels.

“I think what kind of got me hooked on teaching was that the way I learned as a child was not the same as in the classroom,” Morris explains. “And I later saw this in many of my students. I got the feeling that the normal way of learning things was not working for them. And they thought it was their fault. They just didn’t really feel welcome within the traditional education model.”

Morris says that when she worked with those students, tossing aside tradition and instead saying — “You know, we’re just going to do this animation. Or we’re going to make this design or this website or these graphics, and we’re going to approach it in this totally different way” — she saw people “kind of unlock and be like, ‘Oh my gosh. I never thought I could do that.’

“For me, that was the hook, that’s the magic of it. Because I was coming from that experience of having to figure out those unlock mechanisms for myself, it was really exciting to be able to share them with other people, those unlock moments.”

For her doctoral work with the MIT Media Lab’s Fluid Interfaces Group, she’s focusing on the personal space and emotional gaps associated with learning, particularly online and AI-assisted learning. This research builds on her experience increasing human connection in both physical and virtual learning environments.

“I’m developing a framework that combines AI-driven behavioral analysis with human expert assessment to study social learning dynamics,” she says. “My research investigates how social interaction patterns influence curiosity development and intrinsic motivation in learning, with particular focus on understanding how these dynamics differ between real peers and AI-supported environments.”

The first step in her research is determining which elements of social interaction are not replaceable by an AI-based digital tutor. Following that assessment, her goal is to build a prototype platform for experiential learning.

“I’m creating tools that can simultaneously track observable behaviors — like physical actions, language cues, and interaction patterns — while capturing learners’ subjective experiences through reflection and interviews,” Morris explains. “This approach helps connect what people do with how they feel about their learning experience.

“I aim to make two primary contributions: first, analysis tools for studying social learning dynamics; and second, prototype tools that demonstrate practical approaches for supporting social curiosity in digital learning environments. These contributions could help bridge the gap between the efficiency of digital platforms and the rich social interaction that occurs in effective in-person learning.”

Her goals make Morris a perfect fit for the MIT MAD Fellowship. One statement in MAD’s mission is: “Breaking away from traditional education, we foster creativity, critical thinking, making, and collaboration, exploring a range of dynamic approaches to prepare students for complex, real-world challenges.”

Morris wants to help community organizations deal with the rapid AI-powered changes in education, once she finishes her doctorate in 2026. “What should we do with this ‘physical space versus virtual space’ divide?” she asks. That is the space currently captivating Morris’s thoughts.

Research has shown that large language models (LLMs) tend to overemphasize information at the beginning and end of a document or conversation, while neglecting the middle.

This “position bias” means that, if a lawyer is using an LLM-powered virtual assistant to retrieve a certain phrase in a 30-page affidavit, the LLM is more likely to find the right text if it is on the initial or final pages.

MIT researchers have discovered the mechanism behind this phenomenon.

They created a theoretical framework to study how information flows through the machine-learning architecture that forms the backbone of LLMs. They found that certain design choices which control how the model processes input data can cause position bias.

Their experiments revealed that model architectures, particularly those affecting how information is spread across input words within the model, can give rise to or intensify position bias, and that training data also contribute to the problem.

In addition to pinpointing the origins of position bias, their framework can be used to diagnose and correct it in future model designs.

This could lead to more reliable chatbots that stay on topic during long conversations, medical AI systems that reason more fairly when handling a trove of patient data, and code assistants that pay closer attention to all parts of a program.

“These models are black boxes, so as an LLM user, you probably don’t know that position bias can cause your model to be inconsistent. You just feed it your documents in whatever order you want and expect it to work. But by understanding the underlying mechanism of these black-box models better, we can improve them by addressing these limitations,” says Xinyi Wu, a graduate student in the MIT Institute for Data, Systems, and Society (IDSS) and the Laboratory for Information and Decision Systems (LIDS), and first author of a paper on this research.

Her co-authors include Yifei Wang, an MIT postdoc; and senior authors Stefanie Jegelka, an associate professor of electrical engineering and computer science (EECS) and a member of IDSS and the Computer Science and Artificial Intelligence Laboratory (CSAIL); and Ali Jadbabaie, professor and head of the Department of Civil and Environmental Engineering, a core faculty member of IDSS, and a principal investigator in LIDS. The research will be presented at the International Conference on Machine Learning.

Analyzing attention

LLMs like Claude, Llama, and GPT-4 are powered by a type of neural network architecture known as a transformer. Transformers are designed to process sequential data, encoding a sentence into chunks called tokens and then learning the relationships between tokens to predict what words comes next.

These models have gotten very good at this because of the attention mechanism, which uses interconnected layers of data processing nodes to make sense of context by allowing tokens to selectively focus on, or attend to, related tokens.

But if every token can attend to every other token in a 30-page document, that quickly becomes computationally intractable. So, when engineers build transformer models, they often employ attention masking techniques which limit the words a token can attend to.

For instance, a causal mask only allows words to attend to those that came before it.

Engineers also use positional encodings to help the model understand the location of each word in a sentence, improving performance.

The MIT researchers built a graph-based theoretical framework to explore how these modeling choices, attention masks and positional encodings, could affect position bias.

“Everything is coupled and tangled within the attention mechanism, so it is very hard to study. Graphs are a flexible language to describe the dependent relationship among words within the attention mechanism and trace them across multiple layers,” Wu says.

Their theoretical analysis suggested that causal masking gives the model an inherent bias toward the beginning of an input, even when that bias doesn’t exist in the data.

If the earlier words are relatively unimportant for a sentence’s meaning, causal masking can cause the transformer to pay more attention to its beginning anyway.

“While it is often true that earlier words and later words in a sentence are more important, if an LLM is used on a task that is not natural language generation, like ranking or information retrieval, these biases can be extremely harmful,” Wu says.

As a model grows, with additional layers of attention mechanism, this bias is amplified because earlier parts of the input are used more frequently in the model’s reasoning process.

They also found that using positional encodings to link words more strongly to nearby words can mitigate position bias. The technique refocuses the model’s attention in the right place, but its effect can be diluted in models with more attention layers.

And these design choices are only one cause of position bias — some can come from training data the model uses to learn how to prioritize words in a sequence.

“If you know your data are biased in a certain way, then you should also finetune your model on top of adjusting your modeling choices,” Wu says.

Lost in the middle

After they’d established a theoretical framework, the researchers performed experiments in which they systematically varied the position of the correct answer in text sequences for an information retrieval task.

The experiments showed a “lost-in-the-middle” phenomenon, where retrieval accuracy followed a U-shaped pattern. Models performed best if the right answer was located at the beginning of the sequence. Performance declined the closer it got to the middle before rebounding a bit if the correct answer was near the end.

Ultimately, their work suggests that using a different masking technique, removing extra layers from the attention mechanism, or strategically employing positional encodings could reduce position bias and improve a model’s accuracy.

“By doing a combination of theory and experiments, we were able to look at the consequences of model design choices that weren’t clear at the time. If you want to use a model in high-stakes applications, you must know when it will work, when it won’t, and why,” Jadbabaie says.

In the future, the researchers want to further explore the effects of positional encodings and study how position bias could be strategically exploited in certain applications.

“These researchers offer a rare theoretical lens into the attention mechanism at the heart of the transformer model. They provide a compelling analysis that clarifies longstanding quirks in transformer behavior, showing that attention mechanisms, especially with causal masks, inherently bias models toward the beginning of sequences. The paper achieves the best of both worlds — mathematical clarity paired with insights that reach into the guts of real-world systems,” says Amin Saberi, professor and director of the Stanford University Center for Computational Market Design, who was not involved with this work.

This research is supported, in part, by the U.S. Office of Naval Research, the National Science Foundation, and an Alexander von Humboldt Professorship.