Scattering neutrinos to probe the fundamental laws of the universe

Many people who are successful in STEM fields were lucky to have someone who turned them on to a topic and encouraged their interest. For Faith Reyes, that person was her high school physics teacher, Ms. Bolster. 

Whereas Reyes had earlier studied math and science without seeing how those subjects could be applied outside the classroom, her teacher helped her connect those dots and experience the excitement of investigating the physical world. With Ms. Bolster’s help, Reyes started a physics club where students would gather before school and conduct experiments. 

“I just sort of fell in love with physics then,” Reyes says. “And luckily, as I began to study it more and more, I found I landed exactly where I wanted to be.”

Reyes has carried that enthusiasm for physics into not only her research but her role as a personal tutor and teaching assistant at MIT, where she works to foster the same love of the subject in first- and second-year undergraduate students.

As an experimental particle physicist and sixth-year PhD student in the Formaggio Group in the Laboratory for Nuclear Science, Reyes studies neutrinos, elementary particles that have vanishingly little mass and rarely interact with other matter. Neutrinos are produced during radioactive decay, including the processes taking place inside nuclear reactors. Because they interact so infrequently, detecting them can be difficult. (We can’t feel them, but neutrinos from the sun are streaming through our bodies every minute.) Their unusual behavior makes them valuable to physicists seeking to understand what lies beyond the Standard Model, the framework that describes many of the fundamental particles and forces in nature.

“The Standard Model is extremely accurate and describes most of everything that we see,” Reyes says. “But it’s not complete.”

Reyes is a member of the Ricochet neutrino experiment, an international collaboration studying neutrinos produced by a nuclear reactor at the Institut Laue-Langevin in Grenoble, France. The experiment seeks to observe coherent elastic neutrino-nucleus scattering, a low-energy interaction in which a neutrino scatters off an atomic nucleus.

Through Ricochet, scientists aim to investigate some properties of these elusive particles, and contribute to, as Reyes puts it, “just fundamentally understanding the world in which we live.”

When looking back at her time in graduate school, Reyes’ path has not always followed the plan she initially envisioned.

When she joined Ricochet, she expected to work on a particular project located at MIT. But a few years into her PhD, it was clear that the project would not be ready within her timeline. Reyes instead shifted her focus to work taking place in France, where she began learning the technical details of the experiment’s detectors.

Her first visit lasted three months. At the time, Ricochet had two detectors, and Reyes spent much of her time performing the routine work required to understand how they operated.

As the experiment expanded to nine detectors and eventually 18, the amount of work required to manage the system grew substantially. Reyes and a colleague recognized that many of the repetitive tasks could be automated.

Together, they developed a software framework that could perform much of the low-level analysis and detector monitoring that Reyes had initially carried out manually.

The project became an important part of her development as a physicist. By working closely with the detectors and helping build tools to manage them, Reyes gained a detailed understanding of the experiment’s operations.

“It’s sort of like you’re building your own stuff to replace yourself,” she says. “Which is nice in a way because you can save yourself a lot of time.”

The opportunity was both validating and humbling. As a graduate student, she had moved from learning the basics of the experiment to helping guide the work of other scientists.

“It felt like my collaborators trusted me, and I had something of value to give to the collaboration,” Reyes says.

The people she has met through MIT and the Ricochet collaboration have been among the most rewarding parts of her graduate experience. Students, mentors, and collaborators have helped her think critically and become a better physicist, she says.

Her increasing leadership responsibilities have also changed how she approaches research.

Earlier in her academic career, Reyes says, she was more comfortable being told what to do than proposing her own scientific ideas. Over time, leading projects and coordinating groups pushed her to become more confident in her judgment.

“I think I was sometimes not really standing up for myself,” she says. “But now I feel more confident in my position and my prowess as a physicist.”

That confidence has become one of the most important lessons of her PhD.

After completing her doctorate, Reyes hopes to continue conducting research. She is considering a postdoctoral position, which would allow her to continue working in physics at another institution.

Her time in France has also influenced her vision of the future. Reyes spent nine months there through the Chateaubriand Fellowship, following two earlier three-month visits. While the latest trip was primarily focused on research, working in the same office as her collaborators made it easier to coordinate across time zones and strengthened her connection to the experiment. She also grew fond of the country’s culture and work-life balance and would even consider living there. 

“I fell in love with France and the people,” she says. “And also, the work culture.”

Outside the lab, Reyes enjoys playing video games and crocheting, a hobby she picked up during her time in France. She often crochets while watching movies or television, appreciating the opportunity to work with her hands while thinking about other things.

For a physicist whose work involves investigating some of the universe’s smallest and most elusive particles, the hobby offers a different kind of satisfaction: creating something tangible.

As Reyes moves toward the next stage of her career, she hopes to continue pursuing the questions that first drew her to physics. Her research may help reveal what lies beyond the Standard Model, but her experience at MIT has also shown her how much science depends on collaboration, adaptability, and the confidence to lead.

“I’ve learned a lot of lessons,” Reyes says. “Especially about wrangling people.”

de MIT News https://ift.tt/v7O13FE

Estimating suicide risk from text

When people reach out during a mental health crisis, a top priority for counselors is identifying those with a high risk of suicide. The distressed person’s language holds critical clues, and a new tool developed by scientists at MIT’s McGovern Institute for Brain Research is designed to pick up on and rapidly evaluate those signals.

The language-processing tool was developed by Daniel Low, a former graduate student in Senior Research Scientist Satra Ghosh’s Senseable Intelligence Group who is now a research scientist at the Child Mind Institute, where he leads its AI, Risk, and Contemplative Science Lab, as well as a visiting scholar at Harvard University. It uses a custom-built list of words and phrases linked to 49 suicide risk factors, searching text for these and using them to estimate an individual’s risk.

Ghosh, Low, and colleagues report today in the Journal of Psychopathology and Clinical Science that their tool accurately predicts suicide risk from text conversations with crisis counselors. It is already helping to clarify which suicide risk factors matter most in times of crisis. With more validation, it could help with risk assessment in clinical settings and crisis-support situations.

Identifying key risk factors

Suicide attempts are notoriously difficult to predict. Dozens of risk factors have been linked to suicide, and even trained clinicians struggle to identify who will make an attempt among those who have some form of suicidal ideation. Among the factors that can make suicidal thoughts and behaviors more likely are certain psychiatric symptoms and disorders, like depression, borderline personality disorder, and post-traumatic stress disorder, as well as environmental and social stressors, like poverty, incarceration, discrimination, and loneliness.

“You see all these 50 risk factors, and they're all interacting in ways we don't really understand,” Low says. “Many different pathways could lead to someone feeling they want to escape their internal pain,” he says — and it’s challenging to know whose path will lead to a suicide attempt or death.

Ghosh and Low wanted to understand which risk factors counselors and clinicians should most look out for during a mental health crisis. To do that, they collaborated with the Crisis Text Line, whose trained volunteers provide confidential text-based support to people in distress.

Crisis Text Line, a global mental health nonprofit that provides free, 24/7, confidential mental health support for people in need, provided specialized training and controlled access to this restricted dataset. The researchers analyzed de-identified texts from approximately 16,000 conversations with Crisis Text Line’s volunteer crisis counselors. Based on Crisis Text Line’s assessments, those conversations were grouped into three different risk levels: non-suicidal, suicidal ideation without imminent risk, and imminent risk. It was this imminent risk group — those with a plan for suicide, or who have an intent to die within the next 48 hours — that the researchers most wanted to understand.

“We wanted to know what type of symptoms predict the highest suicide risk,” Low says. This question has been studied before, he says — but typically through epidemiological surveys that ask a person to recall their symptoms and experiences, often after their mental health crisis has passed. In contrast, he says, “Crisis Text Line gives us an opportunity to assess many different symptoms and potential risk factors as people are having the crises.”

Reading between the lines

Before analyzing the crisis line texts, the research team built a suicide-risk lexicon. They turned to artificial intelligence to generate a preliminary list of words and phrases tied to established suicide risk factors, including factors associated with suicidal ideation, suicide attempt, and suicide death. Then they manually reviewed and curated that list. Their final lexicon includes about 60 words or phrases for each of 49 risk factors, with the relevance of each one confirmed by expert clinicians.

Then they trained a machine learning model to search the crisis conversations for words and phrases in their lexicon and use these to predict suicide risk. Because the lexicon links each word or phrase to a specific risk factor, they could use these data to determine which risk factors are most closely tied to imminent risk among people in crisis.

What they found was consistent with patterns found in previous research, although not always intuitive. For example, depression is a well-known risk factor for suicidal ideation, but their model found that mentions of lethal means and substance use were more likely to be expressed by the highest-risk group than depressed mood or fatigue. Expressions of active suicidal ideation and self-injury were also strong predictors. Intermediate predictors included anxiety, post-traumatic stress disorder, and emotional pain.

The predictive model assigns a weight to each risk factor based on its contribution to risk. For example, mentions of lethal means for suicide, like “cut” or “pills,” are weighed heavily, whereas terms related to hopelessness, like “don’t know what to do” or “hopeless,” contribute to a lesser degree. After training their model, the team found they could use it to accurately predict risk severity in new conversations the model had not previously seen.

One limitation of lexicons, the researchers note, is that they do not consider the context of terms, and they can miss terms that are similar to those in the lexicon, but not explicitly included. Large language models have reasoning abilities, and Low and colleagues have developed ways of using large language models to detect suicide risk in other projects. However, they say they often use their lexicon in parallel to guarantee flagging certain terms, as well as to maintain data privacy.

Low stresses that while the team used the power of a large language model to develop its lexicon, its prediction model is a simpler, “lightweight” model. Unlike large language models, which require massive computational power, it can be run easily on a personal computer, reducing both cost and privacy concerns. Just as importantly, it is interpretable: Rather than merely generating a risk estimate like some deep learning models can do more effectively, it tells users how it got there. Words of concern can be flagged so users understand the basis for each assessment and act on that information. They are working on similar explainability approaches with large language models.

That’s critical, because the stakes are so high. “This is such a complex space that having a human in the loop is, I think, going to be critical for a long, long time,” says Ghosh, who is the director of the Open Data in Neuroscience Initiative at the McGovern Institute. Likewise, the researchers add that any predictive model must be thoroughly validated before clinical use, and might need to be continually refined to keep up with changes in language use or target populations.

Because a reliable lexicon opens doors to new ways of understanding mental health, Ghosh and Low are widely sharing not just their suicide risk lexicon, but also the software package they developed to build it. Researchers can use that tool to efficiently build lexicons for other mental health conditions. Meanwhile, Low says, the suicide risk lexicon is already being used to explore how text data from a variety of sources, from social media to electronic health records, might help researchers and clinicians better estimate risk.

de MIT News https://ift.tt/8agsxJz

Biologists identify a cellular pathway that allows colorectal cancer to metastasize

Most colon cancer deaths are caused by the spread of tumor cells beyond the colon, usually to the liver. In a new study, MIT biologists identified a cellular pathway necessary for colorectal cancer metastasis.

The pathway they identified, controlled by a protein known as YAP1, is normally involved in tissue repair. When activated in cancer cells, it promotes cell proliferation and migration. The researchers also found that a high-fat diet is more likely to turn on this pathway, through the production of fatty molecules called ceramides.

Drugs that block ceramide production could offer a new way to help prevent metastasis in patients diagnosed with colon cancer, the researchers say.

“We’ve found a pathway that we think is druggable. If we shut down the enzymes that make ceramides, tumor cells can’t switch on this regenerative program, and they largely fail to seed metastases in the liver,” says Omer Yilmaz, director of the MIT Stem Cell Initiative, a professor of biology at MIT and a member of MIT’s Koch Institute for Integrative Cancer Research. He is also a gastrointestinal pathologist and director of translational research in pathology at Beth Israel Deaconess Medical Center.

Yilmaz, Nilay Sethi, an associate professor of medicine at Harvard Medical School and Dana Farber Cancer Institute, and Alpaslan Tasdogan, head of the Institute for Tumor Metabolism and a professor in the Department of Dermatology at University Hospital Essen and the German Cancer Consortium (DKTK), are the senior authors of the study, which appears today in Science. MIT postdocs Swagata Goswami, Qiming Zhang, and Abdullah Burak Yildiz are the paper’s lead authors.

A hijacked pathway

In the United States, colon cancer is usually diagnosed at stage 2 or 3 — before the cancer has spread. However, even after successful surgery, up to a third of these patients will relapse with metastatic disease.

While scientists have identified many genetic mutations that drive the development of colon cancer, it’s unknown exactly what prompts them to spread beyond the colon. 

“Many studies have looked for a genetic driver of metastasis and come up empty,” Yilmaz says. “There isn’t a defining mutational signature that separates metastatic cells from the primary tumor, which points to metastasis being driven largely by changes in which genes are switched on and off, rather than by new mutations.”

In this study, the researchers sought to identify epigenetic programs that enable colon cancer cells to metastasize. Using tumor organoids from mouse models of several types of colon cancer and from patients with colorectal cancer, they found that metastatic cells shared one key feature: activation of the YAP1 program.

YAP1 is a protein that works with partner factors to switch on genes related to development, stem cell maintenance, and regeneration. In normal tissue, it is active during fetal development, and after injury, to promote healing.

In the gut, that repair response runs through a rare, fetal-like cell type, which normally appears only briefly to rebuild the intestinal lining after damage. YAP1 has been linked to cancer for years, but the new work shows that diet-derived lipids push tumor cells into this specific regenerative state — and that the state itself is what licenses metastasis.

“The regenerative program that we described is generally observed in the gut when there is severe injury or infection and the gut needs to regenerate. We see the tumor cells hijack this program to drive metastatic progression,” Goswami says.

Activation of this set of genes helps cancer cells to break free from the original tumor site and spread to other locations in the body. For colon cancer, the most common site of metastasis is the liver, followed by the lungs.

In mouse studies, the researchers also found that cancer cells in animals fed a high-fat diet turned on YAP1 to a greater extent than mice fed a healthy diet. A high-fat diet, the researchers found, triggers activation of enzymes that produce ceramides, a type of lipid. Ceramides then release the molecular brake that normally keeps YAP1 inactive, allowing it to move into the nucleus and switch on its target genes.

Preventing metastasis

The researchers showed that genetically targeting YAP1, or the genes involved in ceramide production, markedly reduced the spread of colon cancer to the liver in mice.

To determine if YAP1 is also involved in metastasis in humans, the researchers analyzed RNA sequencing data from patients with colorectal cancer. They found that YAP1 was more active in metastatic cancer cells, and that patients with higher body mass index (BMI) showed higher expression of the genes activated by YAP1 than normal-weight patients. Patients with higher levels of those genes also had lower survival rates.

“We don’t think that the YAP1 program is specific to obesity. It’s just that it becomes accentuated in obesity, and that may account for why obesity is known to drive the progression of colorectal cancer,” Yilmaz says.

They now plan to develop drugs that inhibit two of the enzymes involved in ceramide production, DEGS1 and DEGS2, in hopes that such drugs could help prevent colon cancer metastasis.

The researchers caution that the findings do not yet translate into dietary advice for patients who have already been diagnosed, and that any drug targeting ceramide synthesis will have to clear a high bar for selectivity, since these lipids are also essential in healthy tissues.

The research was funded by the National Institutes of Health/National Cancer Institute, the MIT Stem Cell Initiative, a Koch Institute Frontier grant, and the NRW Junior Research Program.

de MIT News https://ift.tt/KHyDJsP

New cell-collection device could improve early cancer detection

One of the main reasons that ovarian cancer is among the deadliest forms of cancer is timing: When doctors catch it early, the five-year survival rate can be north of 90 percent. But when doctors catch it late, in stages 3 or 4, five-year survival is less than half that.

About 20 years ago, researchers studying ovarian cancer discovered that many cases of high-grade serous ovarian cancer, the most common type, originate in the fallopian tubes. Detecting the disease there remains challenging, in part because its precursor lesions can be microscopic and difficult to sample.

Now researchers in the group of MIT Professor Kripa Varanasi, working with colleagues at MIT and Johns Hopkins University, have developed a handheld device capable of gently collecting living cells from specific locations to test for ovarian and many other types of cancer. The researchers believe the technique could one day be used to catch cancers earlier and more effectively. It could also be used to create treatments based on individual patient samples.

In a study describing the system in the journal Device, the researchers showed their system enables targeted sampling of newly excised tissue, and they used it to recover living cells for cultivation and testing. The device holds a small microfluidic channel against the tissue and uses a syringe to drive fluid through the channel, applying a force parallel to the tissue surface to gently detach living cells from tiny sections of tissue.

“We wanted to collect living cells from specific regions of the fallopian tube while leaving the surrounding tissue intact,” says Varanasi, senior author of the study and the Maher A. Elmasri Professor of Mechanical Engineering. “Once we have these living cells, there are many things we can do with them. We can use them for diagnostics, grow them into organoids, and build living models of disease. Ultimately, this could allow us to test how an individual patient’s cells respond to different treatments and help us develop more personalized medicines.”

Joining Varanasi on the paper are co-first authors Domitille Avalle SM ’25, MIT postdoc Bert Vandereydt PhD ’26, and Sean Parks ’20, SM ’24. The other authors are MIT PhD candidate Huaiyao Peng; Rebecca Stone, the Johns Hopkins University School of Medicine Stoddard and O’Neil Professor in Gynecologic Oncology; and Angela Belcher, MIT’s James Mason Crafts Professor and a professor of biological engineering and of materials science and engineering.

Living cells for ovarian cancer research

The discovery that many high-grade serous ovarian cancers originate in the fallopian tubes has opened up new prevention options for women at increased risk, who can have their fallopian tubes removed after childbearing years, largely preserving normal hormone production.

Stone, a gynecologic oncologist at Johns Hopkins University, has long advocated for this procedure for certain women at increased risk of ovarian cancer. Belcher introduced Stone to Varanasi, and the three, together with other collaborators, received funding from Break Through Cancer, a foundation that brings together interdisciplinary teams to tackle some of the most challenging problems in cancer. Their project focuses on developing new approaches for the early detection of ovarian cancer, with the cell-collection technology forming one part of that broader effort. 

The researchers began by asking whether they could collect living cells from specific regions of removed fallopian tubes to study the disease’s earliest stages.

"The idea was to see if we could find early signals from precancerous regions of concern,” Varanasi recalls.

The process traditionally involves placing surgically removed fallopian tubes in a chemical preservative and cutting the tissue into sections. The preservative maintains tissue structure, but the cells are no longer alive and cannot be grown in culture. A pathologist then looks for cancerous or precancerous regions in thin sections of the tissue under a microscope. 

“It’s very time-consuming and destructive to the cells,” Varanasi says. “We wanted to bring new capabilities to pathology, so we can not only see what these cells look like, but also collect them alive and study how they behave.”

The MIT researchers saw the process firsthand while visiting surgeons in the operating room at Johns Hopkins.

“It inspired us,” Varanasi says. “We do a lot of work on fluid-surface interfaces in my lab, and we realized we could use a fluid instead of a scalpel or brush, because when you flow a fluid it applies shear stress at the interface. We thought it could work because we heard from surgeons that cells in some locations were loose and would come off during routine washing and other procedures.”

“This is exactly the kind of problem that benefits from bringing clinicians and engineers together,” Stone says. “We understand the clinical need, while the MIT team brings a very different perspective from fluid mechanics and engineering. That combination allowed us to approach the problem in a new way.”

The researchers’ new approach uses a 3D-printed microfluidic device that forms a vacuum seal with the tissue. The device confines liquid flow to a small region, where the flowing liquid creates shear stress that gently detaches living cells.

“We came up with this device where one syringe creates a vacuum that holds it against the tissue, and a second syringe pushes liquid through it,” Vandereydt says. “The vacuum creates a seal, so nothing leaks, and then we locally apply what is basically a microfluidic chip on the tissue that gently shears the cells off.”

The researchers showed they could tune the shear stress applied to the tissue and compared their approach to other cell collection workflows. They found the cells collected using their technique remained viable and grew in culture much more readily than cells detached using conventional approaches.

Finally, the researchers tested their device on fresh human fallopian tube samples, which required them to be on call for sample shipments from their collaborators at Johns Hopkins. After experiments, the samples were shipped back for conventional pathology.

“The samples could come at any time. Sometimes, we’d get an email from our collaborators at 11 p.m. saying ‘There are two fallopian tubes coming tomorrow,’” Vandereydt says. “We were able to collect living cells from those fallopian tubes and turn those into organoids, which is important for testing, disease modeling, and eventually developing personalized treatments.”

“We are developing optical approaches to identify suspicious regions of tissue, and this technology could allow us to collect living cells from exactly those locations,” Belcher says. “Being able to first see where the disease may be emerging and then collect those cells for further study could be very powerful.”

From device to diagnostic

The researchers tested the device on different types of cells and found the approach can be tuned to collect cells of all types by applying different levels of shear stress.

“It’s agnostic to the disease,” Vandereydt says. “There are very loosely adherent prostate cancer cells that detach at 1 pascal [of stress], but if you look at bone cancer cells, only a few cells detach under as high as 5 pascals of stress.” 

The researchers plan for the early use of their device to involve tissue that has already been removed from the body, as that offers an easier pathway to regulatory approval. But they would also like to see their device used to swab samples inside of patients for easier testing and earlier cancer detection.

“What is exciting about this technology is the ability to collect living cells from a specific area while preserving the tissue for pathology,” Stone says. “In the future, one could imagine integrating it into routine histopathology workflows, creating a powerful new way to study carcinogenesis and fundamental biology directly from human tissue.”

Varanasi credits Break Through Cancer for enabling the project.

“Break Through Cancer brought together people working on not only ovarian cancer but also on pancreatic cancer, brain cancer, leukemia, and other cancers,” Varanasi says. “What we heard again and again is how valuable it would be to have better ways to obtain living cells from specific regions of tissue.”

The researchers hope that by making it possible to collect living cells from precise locations without removing or destroying the surrounding tissue, their approach could eventually help researchers and clinicians identify disease earlier and better understand how it develops.

“If this work can ultimately help women by enabling earlier detection of ovarian cancer, I would find that incredibly fulfilling,” Varanasi says. “That is really what motivates us — taking the science and engineering we develop in the lab and using it to make a difference in people’s lives.”

The work was supported by the Break Through Cancer Foundation.

de MIT News https://ift.tt/LJu9e3h

The promise and peril of using visual AI to study cities

A few months ago, researchers from the MIT Senseable City Lab published a study about pollution in New York City featuring some new methods. For instance: With machine learning, they identified the types of vehicles appearing in 331 traffic cameras in the city, and estimated the emissions coming from each automobile. Given enough cameras, these visual artificial intelligence techniques could monitor emissions with an unprecedented combination of precision and scale. 

For that matter, visual AI today can address all kinds of questions for urban planners. Why exactly is traffic snarling? What are the most dangerous aspects of different intersections? Which parts of plazas or parks attract the most people?

Across cities, more images means more data, more insight — and more concerns about privacy and fairness. 

“We can treat these digital images as data and quantify features of the city,” says Fábio Duarte, an MIT researcher and co-author of a new book about visual AI and urban studies. “With computer vision techniques, each image is a dataset.” Still, he adds, “We have to be careful about it.” 

And while urbanists have long used visual analysis to inform their thinking, now it’s possible to an unprecedented extent. 

“Everybody has been observing the urban environment and trying to get some insight,” says Martina Mazzarello, an MIT scholar and a co-author of the new book. “But what if we can do that at a large scale and get some insight everywhere?”

The scholars explore these topics in “How AI Sees the City: Urban Visual Intelligence,” published this month by Routledge. The authors are Duarte, a principal research scientist and associate director of the MIT Senseable City Lab; Mazzarello, a research scientist and lead of MIT Senseable City Lab global initiatives; Carlo Ratti, a professor of the practice and founder and director of the MIT Senseable City Lab; and Fan Zhang, an assistant professor at the Institute of Remote Sensing and GIS at Peking University.

“Great urbanists such as Kevin Lynch and Willian H. Whyte showed us the extraordinary value of ‘looking’ at the city,” Ratti says, referring to two prominent thinkers about city dynamics whose work is described in the book. “Today, visual AI gives us new ways to build on that tradition — allowing us to observe cities at a scale and with a level of detail that was previously impossible.” 

New tool, long tradition

“How AI Sees the City” stems from the work of the MIT Senseable City Lab, founded in 2004, which uses data to better understand urban dynamics. As the authors discuss in the book, there is a long history of visual representations that shape the way we think about cities, from Romans building marble maps to the introduction of photography — which produced influential urban images about things like Hausmann’s reshaping of Paris or the crowding of tenements in New York City’s Lower East Side during the 19th century.

More recently, some scholars have used visual studies to better understand city form, including Lynch, a former MIT professor whose 1960 book, “The Image of the City,” influenced many scholars. Whyte, a sociologist famous for his book “The Organization Man,” then became an urbanist closely examining public spaces.

By explicitly placing AI in a continuum with these visual urban studies, the authors are making a point: Powerful as it might be, we can still think of AI primarily as a tool serving human purposes, as we seek to design and refine urban form.

“Kevin Lynch at MIT was only using paper and pen,” Duarte says. “We can now scale up what he was doing, with visual AI, while also looking at many different dimension of cities.” 

There are extensive possibilities for applying visual AI to urban planning, ranging from emissions to traffic flow, safety, better imagery of street-level activity and sidewalks, and much more. The book also examines, for instance, urban greenery. While satellite imagery can show us how much tree cover and green spaces cities have, near-ubiquitous images from phones and other sources can also reveal to what extent people glimpse greenery in everyday life, a factor in reported wellness.

“The real promise of visual AI is not simply that computers can look at millions of images,” Zhang says. “It is that we can connect what is visible in those images — streets, buildings, greenery, traffic, public space — with larger questions about how cities function and how people experience them.”

Better image recognition by AI even extends to urban interiors. By using images from 400,000 AirBnB listings across the world, one recent Senseable City study shows that, contrary to some claims, interior design styles are not becoming globally more homogeneous, but reflect significant geographic differences. 

“No matter what it is, we can learn from what we can see and then use it as urban designers, planners, policymakers, and citizens,” Mazzarello says. “It can be our eyes, or cameras with computers, but in the end it’s the same methodology, and now we are trying to optimize the ways we can use these tools.”

Promise and pitfalls

If the promise of visual AI for urban studies is vast, the pitfalls are concerning. In “How AI Sees the City,” the authors outline multiple potential problems with the technology, including the intrusiveness of widespread visual surveillance and the potential for bias being reinforced through AI systems. 

The installation of ubiquitous cameras can quickly raise concerns about surveillance. London, an early adopter of CCTV, has about 210 cameras per square mile. But eight of the world’s 10 most camera-heavy cities are in China; Shanghai has over 5,000 cameras per square mile. Such surveillance practices have raised controversy in other parts of the world, with debate over the uses of traffic cameras bubbling up in the U.S. this year as well. 

In evaluating the potential safety gains from intensive video recording, the authors write, “the benefits must be weighed against the significant erosion of personal freedom and the potential for abuse inherent in a system of constant monitoring.”

Meanwhile, AI systems can reinforce social biases as well, leading to the production of data that reinforce prior perceptions as much as underlying realities — about people, neighborhoods, and whole cities. If AI models are trained on majority population groups, they may not evaluate minority groups the same way. 

“We need to teach AI to see, and depending on how you teach it, it will see what what is embedded in the culture,” Duarte says. “AI is not neutral.”

Still, as Mazzarello adds, “our eyes are not neutral, either. Every tool has to be guided in the right way, and trained in the best way.” 

Other scholars have praised “How AI Sees the City.” Michael Batty of University College London has called it a “fascinating book” that “shows how we are beginning to interpret the world of urban design, suggesting ways in which we might improve design using urban analytics, AI and large language models.”

Ultimately, though the authors think there is great value in deploying visual AI to learn more about our cities, how they function, and how they might be improved. With caution and independent thinking, progress is possible. Or, as they conclude in the book, “We should explore this wisely, critically, and creatively.” 

de MIT News https://ift.tt/OKlAiS3

MIT welcomes David Siegel SM ’86, PhD ’91 as its next Innovation Fellow

David Siegel SM ’86, PhD ’91, a computer scientist, entrepreneur, and philanthropist, will serve as the next MIT Innovation Fellow during the 2026-27 academic year. Working with the MIT Schwarzman College of Computing, Siegel will explore how artificial intelligence can accelerate scientific discovery at the Institute and beyond.

“From the MIT Schwarzman College of Computing to the MIT Siegel Family Quest for Intelligence, David has been a superb thought partner for me and other Institute leaders on a range of very significant initiatives, so we're delighted to have him join us now as an MIT Innovation Fellow,” says MIT President Sally Kornbluth. “Our community has long benefited from David's exceptional technical insight, entrepreneurial experience, instinct for connecting people, and infectious love for MIT. We look forward to working with him now as he helps us identify new opportunities at the intersection of AI and scientific discovery.” 

“MIT has played a foundational role in shaping how I view technology’s potential to address complex challenges,” says Siegel. “I’m thrilled to return to campus as an Innovation Fellow to collaborate with brilliant students, researchers, and faculty at a pivotal juncture in how technology shapes our world.”

A long-standing connection to MIT

Siegel’s relationship with MIT began during his graduate studies, where he earned a master’s degree in 1986 and PhD in 1991, after earning his bachelor’s degree in electrical engineering and computer science from Princeton University in 1983. Immersed in the field during a foundational era for computer science, he worked at the MIT Artificial Intelligence Lab (now the Computer Science and Artificial Intelligence Laboratory) in Professor Tomás Lozano-Pérez’s research group on human-machine interaction, contributing to the development of a pioneering humanlike robotic hand. 

Siegel has remained deeply involved with the Institute in the years since. He is a life member of the MIT Corporation, previously served on its Executive Committee, and co-chairs the External Advisory Committee for the MIT Schwarzman College of Computing. Additionally, Siegel was an early champion of the MIT Quest for Intelligence, an Institute-wide initiative studying intelligence in brains and machines, recently renamed the MIT Siegel Family Quest for Intelligence.

Entrepreneurship, philanthropy, and AI leadership

After completing his PhD, Siegel founded several early internet ventures before co-founding Two Sigma in 2001. A leading global investment firm, Two Sigma approaches investing through a data science and engineering lens, echoing Siegel’s experience in the MIT AI Lab. With the scientific method embedded in its culture, the firm uses artificial intelligence, machine learning, and advanced quantitative modeling. Siegel retired from day-to-day management of Two Sigma in 2024; however, he remains co-chair. 

Currently, Siegel’s work spans several fields, with a strong emphasis on science, technology, and philanthropy. Through the Siegel Family Endowment, a philanthropic foundation that he established in 2011, Siegel supports leaders, researchers, and organizations that are examining how technological change affects society and how to guide that shift for the public good. The endowment backs organizations such as the Scratch Foundation, Center on Rural Innovation, Khan Academy, Pursuit, and The Aspen Institute.

Recognizing the critical resource gap between academic research labs and frontier AI, Siegel founded the nonprofit Open Athena in 2024. Open Athena equips academic research labs with elite AI talent, data engineering expertise, and computational resources to enable groundbreaking discoveries at scale. The organization is also developing Marin, a 535-billion-parameter foundation model built entirely in public. By sharing every dataset and experiment in real-time, Marin ensures that the science of frontier AI remains a shared public asset for researchers and innovators worldwide. Open Athena works with leading global institutions including MIT and is funded by philanthropic partners including Bloomberg Philanthropies, Google, The Huang Foundation, and Schmidt Sciences. 

Siegel actively serves on several governance and advisory boards. He is vice-chair of the Scratch Foundation, which he co-founded in 2013 with MIT Professor Mitch Resnick, a member of the Cornell Tech Council, and a board member of organizations such as Re:Build Manufacturing, Khan Academy, and NYC FIRST Robotics. In 2025, Siegel was appointed to the U.S. Department of Energy’s Office of Science Advisory Committee, providing counsel on complex scientific and technical issues impacting federal scientific research programs.

Outside of philanthropy, Siegel remains actively engaged in the global AI ecosystem as an investor, hands-on advisor, and thought leader. Through his family office Shinrai Management, he focuses on supporting entrepreneurs and investing in high-growth startups, including several founded by MIT students and alumni.

Siegel’s debut book, “When Machines Act: The Promise and Peril of Navigating Our Agentic AI Future,” co-authored with Yale University’s Jeffrey Sonnenfeld and Stephen Henriques, will be published by MIT Press next March, coinciding with his residency as an MIT Innovation Fellow. Drawing on interviews with top tech leaders and off-the-record discussions with over 300 CEOs, the book provides a practical roadmap for autonomous AI, outlining where to deploy it, how to govern it, and which rules truly matter.

“David’s ties to MIT date back to his doctoral research in the AI Lab and have deepened through his many contributions to the Institute, including his significant involvement with the Quest for Intelligence and the MIT Schwarzman College of Computing,” says MIT Provost Anantha Chandrakasan. “His long-standing commitment to MIT, together with his vision for the future of AI and science, makes him an especially fitting Innovation Fellow. We look forward to the contributions he will make and the connections his work will foster across campus."

“For the MIT Schwarzman College of Computing, David’s fellowship is a chance to build on an already strong connection and explore how AI can expand the frontiers of science. Having known David since we were both graduate students at MIT, I’m deeply familiar with his ability to advance AI and its application in various fields,” says Dan Huttenlocher, dean of the MIT Schwarzman College of Computing and the Panasonic Professor of Electrical Engineering and Computer Science. “He understands both the college’s aspirations and the challenges ahead, and his perspective will help us identify concrete paths for research, education, and broader engagement. I look forward to working with him over the coming year.”

A year in residence

MIT Innovation Fellows typically spend a year or more in residence at the Institute. They draw on their experience, expertise, and professional networks to engage with faculty and students, participate in public events, and provide strategic counsel to MIT leaders.

The program has brought luminaries from industry and government to MIT. Most recently, Brian Deese, former White House National Economic Council director, served as an Innovation Fellow. Other fellows have included Virginia M. “Ginny” Rometty, former chair, president, and CEO of IBM; Eric Schmidt, former executive chair of Google’s parent company, Alphabet; the late Ash Carter, former U.S. secretary of defense; and former Massachusetts Governor Deval Patrick.

As an Innovation Fellow, Siegel will help guide the Institute’s focus on leveraging artificial intelligence to support scientific discovery, working closely with the MIT Schwarzman College of Computing and departments across the college to help extend their impact beyond MIT. 

“Using AI to accelerate scientific discovery is the ultimate engineering challenge. There is simply no better launchpad in the world for that work than MIT,” says Siegel.

de MIT News https://ift.tt/5Cv03Dj

FUNdaMENTALs of precision design

Repeatability in engineering product design ensures that a manufacturing process performs the same way every time and allows for a working prototype to be transformed into a reliable, safe, consistent, and cost-effective mass-market product. For students in class 2.70 (Fundamentals of Precision Product Design), precision and repeatability are the name of the game. 

“[As an engineer], you have an extra responsibility to overlook nothing,” says course instructor Alex Slocum, the Walter M. and A. Hazel May Professor of Mechanical Engineering. “If you miss something, someone could be hurt or die.”

Slocum’s message is serious, but his approach to teaching the material is famously fun — in fact, he prefers the spelling “FUNdaMENTALs” for the first word of the class name. His mother, Mariana Polonsky Slocum, was a mathematics professor at MIT. “She taught me, physics doesn't care about your feelings,” he says. “I want [students] to understand that we are governed by the laws of physics, and that is a catalyst for creativity, not a hindrance. It is a hindrance if you forget that.” 

Through the course, students learn deterministic design, selection, and assembly of machine elements to create and manufacture robust precision machines, instruments, and systems. They also apply Slocum’s “Functional Requirements, Ergonomics and Environment, Design Parameters, Analysis, References, Risks, Countermeasures” (FRED PARRC, pronounced like “Fred Park”) model, and engage in peer review and evaluation.

“You get a lot of time working on problems that just pop up in engineering. To me, it felt very [representative] of the grad work that I was doing,” says Mariia Smyk, a graduate student in mechanical engineering. 

Some students may describe the course as “creative chaos,” but tend to agree that their learning experience is one that drives home the fundamentals. 

“It definitely made me more confident knowing that I can look at what I'm designing and be very deliberate in taking steps toward mitigating the risks that anyone would face when they use a product,” says graduate student Adian Salazar. “I feel like I've been able to apply all those really fundamental concepts that I learned in the more theory-heavy classes to real-world machines.”

de MIT News https://ift.tt/fv7ZOlX