MIT welcomes Frida Polli as its next visiting innovation scholar

Frida Polli, a neuroscientist, entrepreneur, investor, and inventor known for her leading-edge contributions at the crossroads of behavioral science and artificial intelligence, is MIT’s new visiting innovation scholar for the 2024-25 academic year. She is the first visiting innovation scholar to be housed within the MIT Schwarzman College of Computing.

Polli began her career in academic neuroscience with a focus on multimodal brain imaging related to health and disease. She was a fellow at the Psychiatric Neuroimaging Group at Mass General Brigham and Harvard Medical School. She then joined the Department of Brain and Cognitive Sciences at MIT as a postdoc, where she worked with John Gabrieli, the Grover Hermann Professor of Health Sciences and Technology and a professor of brain and cognitive sciences.

Her research has won many awards, including a Young Investigator Award from the Brain and Behavior Research Foundation. She authored over 30 peer-reviewed articles, with notable publications in the Proceedings of the National Academy of Sciences, the Journal of Neuroscience, and Brain. She transitioned from academia to entrepreneurship by completing her MBA at the Harvard Business School (HBS) as a Robert Kaplan Life Science Fellow. During this time, she also won the Life Sciences Track and the Audience Choice Award in the 2010 MIT $100K Entrepreneurship competition as a member of Aukera Therapeutics.

After HBS, Polli launched pymetrics, which harnessed advancements in cognitive science and machine learning to develop analytics-driven decision-making and performance enhancement software for the human capital sector. She holds multiple patents for the technology developed at pymetrics, which she co-founded in 2012 and led as CEO until her successful exit in 2022. Pymetrics was a World Economic Forum’s Technology Pioneer and Global Innovator, an Inc. 5000’s Fastest-Growing company, and Forbes Artificial Intelligence 50 company. Polli and pymetrics also played a pivotal role in passing the first-in-the-nation algorithmic bias law — New York’s Automated Employment Decision Tool law — which went into effect in July 2023.

Making her return to MIT as a visiting innovation scholar, Polli is collaborating closely with Sendhil Mullainathan, the Peter de Florez Professor in the departments of Electrical Engineering and Computer Science and Economics, and a principal investigator in the Laboratory for Information and Decision Systems. With Mullainathan, she is working to bring together a broad array of faculty, students, and postdocs across MIT to address concrete problems where humans and algorithms intersect, to develop a new subdomain of computer science specific to behavioral science, and to train the next generation of scientists to be bilingual in these two fields.

“Sometimes you get lucky, and sometimes you get unreasonably lucky. Frida has thrived in each of the facets we’re looking to have impact in — academia, civil society, and the marketplace. She combines a startup mentality with an abiding interest in positive social impact, while capable of ensuring the kind of intellectual rigor MIT demands. It’s an exceptionally rare combination, one we are unreasonably lucky to have,” says Mullainathan.

“People are increasingly interacting with algorithms, often with poor results, because most algorithms are not built with human interplay in mind,” says Polli. “We will focus on designing algorithms that will work synergistically with people. Only such algorithms can help us address large societal challenges in education, health care, poverty, et cetera.”

Polli was recognized as one of Inc.'s Top 100 Female Founders in 2019, followed by being named to Entrepreneur's Top 100 Powerful Women in 2020, and to the 2024 list of 100 Brilliant Women in AI Ethics. Her work has been highlighted by major outlets including The New York Times, The Wall Street Journal, The Financial Times, The Economist, Fortune, Harvard Business Review, Fast Company, Bloomberg, and Inc.

Beyond her role at pymetrics, she founded Alethia AI in 2023, an organization focused on promoting transparency in technology, and in 2024, she launched Rosalind Ventures, dedicated to investing in women founders in science and health care. She is also an advisor at the Buck Institute’s Center for Healthy Aging in Women.

"I'm delighted to welcome Dr. Polli back to MIT. As a bilingual expert in both behavioral science and AI, she is a natural fit for the college. Her entrepreneurial background makes her a terrific inaugural visiting innovation scholar,” says Dan Huttenlocher, dean of the MIT Schwarzman College of Computing and the Henry Ellis Warren Professor of Electrical Engineering and Computer Science.

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

Need a research hypothesis? Ask AI.

Crafting a unique and promising research hypothesis is a fundamental skill for any scientist. It can also be time consuming: New PhD candidates might spend the first year of their program trying to decide exactly what to explore in their experiments. What if artificial intelligence could help?

MIT researchers have created a way to autonomously generate and evaluate promising research hypotheses across fields, through human-AI collaboration. In a new paper, they describe how they used this framework to create evidence-driven hypotheses that align with unmet research needs in the field of biologically inspired materials.

Published Wednesday in Advanced Materials, the study was co-authored by Alireza Ghafarollahi, a postdoc in the Laboratory for Atomistic and Molecular Mechanics (LAMM), and Markus Buehler, the Jerry McAfee Professor in Engineering in MIT’s departments of Civil and Environmental Engineering and of Mechanical Engineering and director of LAMM.

The framework, which the researchers call SciAgents, consists of multiple AI agents, each with specific capabilities and access to data, that leverage “graph reasoning” methods, where AI models utilize a knowledge graph that organizes and defines relationships between diverse scientific concepts. The multi-agent approach mimics the way biological systems organize themselves as groups of elementary building blocks. Buehler notes that this “divide and conquer” principle is a prominent paradigm in biology at many levels, from materials to swarms of insects to civilizations — all examples where the total intelligence is much greater than the sum of individuals’ abilities.

“By using multiple AI agents, we’re trying to simulate the process by which communities of scientists make discoveries,” says Buehler. “At MIT, we do that by having a bunch of people with different backgrounds working together and bumping into each other at coffee shops or in MIT’s Infinite Corridor. But that's very coincidental and slow. Our quest is to simulate the process of discovery by exploring whether AI systems can be creative and make discoveries.”

Automating good ideas

As recent developments have demonstrated, large language models (LLMs) have shown an impressive ability to answer questions, summarize information, and execute simple tasks. But they are quite limited when it comes to generating new ideas from scratch. The MIT researchers wanted to design a system that enabled AI models to perform a more sophisticated, multistep process that goes beyond recalling information learned during training, to extrapolate and create new knowledge.

The foundation of their approach is an ontological knowledge graph, which organizes and makes connections between diverse scientific concepts. To make the graphs, the researchers feed a set of scientific papers into a generative AI model. In previous work, Buehler used a field of math known as category theory to help the AI model develop abstractions of scientific concepts as graphs, rooted in defining relationships between components, in a way that could be analyzed by other models through a process called graph reasoning. This focuses AI models on developing a more principled way to understand concepts; it also allows them to generalize better across domains.

“This is really important for us to create science-focused AI models, as scientific theories are typically rooted in generalizable principles rather than just knowledge recall,” Buehler says. “By focusing AI models on ‘thinking’ in such a manner, we can leapfrog beyond conventional methods and explore more creative uses of AI.”

For the most recent paper, the researchers used about 1,000 scientific studies on biological materials, but Buehler says the knowledge graphs could be generated using far more or fewer research papers from any field.

With the graph established, the researchers developed an AI system for scientific discovery, with multiple models specialized to play specific roles in the system. Most of the components were built off of OpenAI’s ChatGPT-4 series models and made use of a technique known as in-context learning, in which prompts provide contextual information about the model’s role in the system while allowing it to learn from data provided.

The individual agents in the framework interact with each other to collectively solve a complex problem that none of them would be able to do alone. The first task they are given is to generate the research hypothesis. The LLM interactions start after a subgraph has been defined from the knowledge graph, which can happen randomly or by manually entering a pair of keywords discussed in the papers.

In the framework, a language model the researchers named the “Ontologist” is tasked with defining scientific terms in the papers and examining the connections between them, fleshing out the knowledge graph. A model named “Scientist 1” then crafts a research proposal based on factors like its ability to uncover unexpected properties and novelty. The proposal includes a discussion of potential findings, the impact of the research, and a guess at the underlying mechanisms of action. A “Scientist 2” model expands on the idea, suggesting specific experimental and simulation approaches and making other improvements. Finally, a “Critic” model highlights its strengths and weaknesses and suggests further improvements.

“It’s about building a team of experts that are not all thinking the same way,” Buehler says. “They have to think differently and have different capabilities. The Critic agent is deliberately programmed to critique the others, so you don't have everybody agreeing and saying it’s a great idea. You have an agent saying, ‘There’s a weakness here, can you explain it better?’ That makes the output much different from single models.”

Other agents in the system are able to search existing literature, which provides the system with a way to not only assess feasibility but also create and assess the novelty of each idea.

Making the system stronger

To validate their approach, Buehler and Ghafarollahi built a knowledge graph based on the words “silk” and “energy intensive.” Using the framework, the “Scientist 1” model proposed integrating silk with dandelion-based pigments to create biomaterials with enhanced optical and mechanical properties. The model predicted the material would be significantly stronger than traditional silk materials and require less energy to process.

Scientist 2 then made suggestions, such as using specific molecular dynamic simulation tools to explore how the proposed materials would interact, adding that a good application for the material would be a bioinspired adhesive. The Critic model then highlighted several strengths of the proposed material and areas for improvement, such as its scalability, long-term stability, and the environmental impacts of solvent use. To address those concerns, the Critic suggested conducting pilot studies for process validation and performing rigorous analyses of material durability.

The researchers also conducted other experiments with randomly chosen keywords, which produced various original hypotheses about more efficient biomimetic microfluidic chips, enhancing the mechanical properties of collagen-based scaffolds, and the interaction between graphene and amyloid fibrils to create bioelectronic devices.

“The system was able to come up with these new, rigorous ideas based on the path from the knowledge graph,” Ghafarollahi says. “In terms of novelty and applicability, the materials seemed robust and novel. In future work, we’re going to generate thousands, or tens of thousands, of new research ideas, and then we can categorize them, try to understand better how these materials are generated and how they could be improved further.”

Going forward, the researchers hope to incorporate new tools for retrieving information and running simulations into their frameworks. They can also easily swap out the foundation models in their frameworks for more advanced models, allowing the system to adapt with the latest innovations in AI.

“Because of the way these agents interact, an improvement in one model, even if it’s slight, has a huge impact on the overall behaviors and output of the system,” Buehler says.

Since releasing a preprint with open-source details of their approach, the researchers have been contacted by hundreds of people interested in using the frameworks in diverse scientific fields and even areas like finance and cybersecurity.

“There’s a lot of stuff you can do without having to go to the lab,” Buehler says. “You want to basically go to the lab at the very end of the process. The lab is expensive and takes a long time, so you want a system that can drill very deep into the best ideas, formulating the best hypotheses and accurately predicting emergent behaviors. Our vision is to make this easy to use, so you can use an app to bring in other ideas or drag in datasets to really challenge the model to make new discoveries.”

de MIT News https://ift.tt/PwFh62v

How humans continuously adapt while walking stably

Researchers have developed a model that explains how humans adapt continuously during complex tasks, like walking, while remaining stable.

The findings were detailed in a recent paper published in the journal Nature Communications authored by Nidhi Seethapathi, an assistant professor in MIT’s Department of Brain and Cognitive Sciences; Barrett C. Clark, a robotics software engineer at Bright Minds Inc.; and Manoj Srinivasan, an associate professor in the Department of Mechanical and Aerospace Engineering at Ohio State University.

In episodic tasks, like reaching for an object, errors during one episode do not affect the next episode. In tasks like locomotion, errors can have a cascade of short-term and long-term consequences to stability unless they are controlled. This makes the challenge of adapting locomotion in a new environment  more complex.

"Much of our prior theoretical understanding of adaptation has been limited to episodic tasks, such as reaching for an object in a novel environment," Seethapathi says. "This new theoretical model captures adaptation phenomena in continuous long-horizon tasks in multiple locomotor settings."

To build the model, the researchers identified general principles of locomotor adaptation across a variety of task settings, and  developed a unified modular and hierarchical model of locomotor adaptation, with each component having its own unique mathematical structure.

The resulting model successfully encapsulates how humans adapt their walking in novel settings such as on a split-belt treadmill with each foot at a different speed, wearing asymmetric leg weights, and wearing  an exoskeleton. The authors report that the model successfully reproduced human locomotor adaptation phenomena across novel settings in 10 prior studies and correctly predicted the adaptation behavior observed in two new experiments conducted as part of the study.

The model has potential applications in sensorimotor learning, rehabilitation, and wearable robotics.

"Having a model that can predict how a person will adapt to a new environment has immense utility for engineering better rehabilitation paradigms and wearable robot control," Seethapathi says. "You can think of a wearable robot itself as a new environment for the person to move in, and our model can be used to predict how a person will adapt for different robot settings. Understanding such human-robot adaptation is currently an experimentally intensive process, and our model  could help speed up the process by narrowing the search space."

de MIT News https://ift.tt/dD6B1cj

Turning adversity into opportunity

Sujood Eldouma always knew she loved math; she just didn’t know how to use it for good in the world. 

But after a personal and educational journey that took her from Sudan to Cairo to London, all while leveraging MIT Open Learning’s online educational resources, she finally knows the answer: data science.

An early love of data

Eldouma grew up in Omdurman, Sudan, with her parents and siblings. She always had an affinity for STEM subjects, and at the University of Khartoum she majored in electrical and electronic engineering with a focus in control and instrumentation engineering.

In her second year at university, Eldouma struggled with her first coding courses in C++ and C#, which are general-purpose programming languages. When a teaching assistant introduced Eldouma and her classmates to MIT OpenCourseWare for additional support, she promptly worked through OpenCourseWare’s C++ and C courses in tandem with her in-person classes. This began Eldouma’s ongoing connection with the open educational resources available through MIT Open Learning.

OpenCourseWare, part of MIT Open Learning, offers a free collection of materials from thousands of MIT courses, spanning the entire curriculum. To date, Eldouma has explored over 20 OpenCourseWare courses, and she says it is a resource she returns to regularly.

“We started watching the videos and reading the materials, and it made our lives easier,” says Eldouma. “I took many OpenCourseWare courses in parallel with my classes throughout my undergrad, because we still did the same material. OpenCourseWare courses are structured differently and have different resources and textbooks, but at the end of the day it’s the same content.”

For her graduation thesis, Eldouma did a project on disaster response and management in complex contexts, because at the time, Sudan was suffering from heavy floods and the country had limited resources to respond.

“That’s when I realized I really love data, and I wanted to explore that more,” she says.

While Eldouma loves math, she always wanted to find ways to use it for good. Through the early exposure to data science and statistical methods at her university, she saw how data science leverages math for real-world impact.

After graduation, she took a job at the DAL Group, the largest Sudanese conglomerate, where she helped to incorporate data science and new technologies to automate processes within the company. When civil war erupted in Sudan in April 2023, life as Eldouma knew it was turned upside down, and her family was forced to make the difficult choice to relocate to Egypt.

Purpose in adversity

Soon after relocating to Egypt, Eldouma lost her job and found herself struggling to find purpose in the life circumstances she had been handed. Due to visa restrictions, challenges getting right-to-work permits, and a complicated employment market in Egypt, she was also unable to find a new job.

“I was sort of in a depressive episode, because of all that was happening,” she reflects. “It just hit me that I lost everything that I know, everything that I love. I’m in a new country. I need to start from scratch.”

Around this time, a friend who knew Eldouma was curious about data science sent her the link to apply to the MIT Emerging Talent Certificate in Data and Computer Science. With less than 24 hours before the application deadline, Eldouma hit “Submit.”

Finding community and joy through learning

Part of MIT Open Learning, MIT Emerging Talent at the MIT Jameel World Education Lab (J-WEL) develops global education programs that target the needs of talented individuals from challenging economic and social circumstances by equipping them with the knowledge and tools to advance their education and careers.

The Certificate in Computer and Data Science is a year-long online learning program that follows an agile continuous education model. It incorporates computer science and data analysis coursework from MITx, professional skill building, experiential learning, apprenticeship options, and opportunities for networking with MIT’s global community. The program is targeted toward refugees, migrants, and first-generation low-income students from historically marginalized backgrounds and underserved communities worldwide.

Although Eldouma had used data science in her role at the DAL Group, she was happy to have a proper introduction to the field and to find joy in learning again. She also found community, support, and inspiration from her classmates who were connected to each other not just by their academic pursuits, but by their shared life challenges. The cohort of 100 students stayed in close contact through the program, both for casual conversation and for group work.

“In the final step of the Emerging Talent program, learners apply their computer and data knowledge in an experiential learning opportunity,” says Megan Mitchell, associate director for Pathways for Talent and acting director of J-WEL. “The experiential learning opportunity takes the form of an internship, apprenticeship, or an independent or collaborative project, and allows students to apply their knowledge in real-world settings and build practical skills.”

Determined to apply her newly acquired knowledge in a meaningful way, Eldouma and fellow displaced Sudanese classmates designed a project to help solve a problem in their home country. The group identified access to education as a major problem facing Sudanese people, with schooling disrupted due to the conflict. Focusing on the higher education audience, the group partnered with community platform Nas Al Sudan to create a centralized database where students can search for scholarships and other opportunities to continue their education.

Eldouma completed the MIT Emerging Talent program in June 2024 with a clear vision to pursue a career in data science, and the confidence to achieve that goal. In fact, she had already taken the steps to get there: halfway through the certificate program, she applied and was accepted to the MITx MicroMasters program in Statistics and Data Science at Open Learning and the London School of Economics (LSE) Masters of Science in Data Science.

In January 2024, Eldouma started the MicroMasters program with 12 of her Emerging Talent peers. While the MIT Emerging Talent program is focused on undergraduate-level, introductory computer and data science material, the MicroMasters program in Statistics and Data Science is graduate-level learning. MicroMasters programs are a series of courses that provide deep learning in a specific career field, and learners that successfully earn the credential may receive academic credit to universities around the world. This makes the credential a pathway to over 50 master’s degree programs and other advanced degrees, including at MIT. Eldouma believes that her experience in the MicroMasters courses prepared her well for the expectations of the LSE program.

After finishing the MicroMasters and LSE programs, Eldouma aspires to a career using data science to better understand what is happening on the African continent from an economic and social point of view. She hopes to contribute to solutions to conflicts across the region. And, someday, she hopes to move back to Sudan.

“My family’s roots are there. I have memories there,” she says. “I miss walking in the street and the background noise is the same language that I am thinking in. I don’t think I will ever find that in any place like Sudan.”

de MIT News https://ift.tt/1C9qgAV

Miracle, or marginal gain?

From 1960 to 1989, South Korea experienced a famous economic boom, with real GDP per capita growing by an annual average of 6.82 percent. Many observers have attributed this to industrial policy, the practice of giving government support to specific industrial sectors. In this case, industrial policy is often thought to have powered a generation of growth.

Did it, though? An innovative study by four scholars, including two MIT economists, suggests that overall GDP growth attributable to industrial policy is relatively limited. Using global trade data to evaluate changes in industrial capacity within countries, the research finds that industrial policy raises long-run GDP by only 1.08 percent in generally favorable circumstances, and up to 4.06 percent if additional factors are aligned — a distinctly smaller gain than an annually compounding rate of 6.82 percent.

The study is meaningful not just because of the bottom-line numbers, but for the reasons behind them. The research indicates, for instance, that local consumer demand can curb the impact of industrial policy. Even when a country alters its output, demand for those goods may not shift as extensively, putting a ceiling on directed growth.

“In most cases, the gains are not going to be enormous,” says MIT economist Arnaud Costinot, co-author of a new paper detailing the research. “They are there, but in terms of magnitude, the gains are nowhere near the full scope of the South Korean experience, which is the poster child for an industrial policy success story.”

The research combines empirical data and economic theory, using data to assess “textbook” conditions where industrial policy would seem most merited.

“Many think that, for countries like China, Japan, and other East Asian giants, and perhaps even the U.S., some form of industrial policy played a big role in their success stories,” says Dave Donaldson, an MIT economist and another co-author of the paper. “The question is whether the textbook argument for industrial policy fully explains those successes, and our punchline would be, no, we don’t think it can.”

The paper, “The Textbook Case for Industrial Policy: Theory Meets Data,” appears in the Journal of Political Economy. The authors are Dominick Bartelme, an independent researcher; Costinot, the Ford Professor of Economics in MIT’s Department of Economics; Donaldson, the Class of 1949 Professor of Economics in MIT’s Department of Economics; and Andres Rodriguez-Clare, the Edward G. and Nancy S. Jordan Professor of Economics at the University of California at Berkeley.

Reverse-engineering new insights

Opponents of industrial policy have long advocated for a more market-centered approach to economics. And yet, over the last several decades globally, even where political leaders publicly back a laissez-faire approach, many governments have still found reasons to support particular industries. Beyond that, people have long cited East Asia’s economic rise as a point in favor of industrial policy.

The scholars say the “textbook case” for industrial policy is a scenario where some economic sectors are subject to external economies of scale but others are not.

That means firms within an industry have an external effect on the productivity of other firms in that same industry, which could happen via the spread of knowledge.

If an industry becomes both bigger and more productive, it may make cheaper goods that can be exported more competitively. The study is based on the insight that global trade statistics can tell us something important about the changes in industry-specific capacities within countries. That — combined with other metrics about national economies — allows the economists to scrutinize the overall gains deriving from those changes and to assess the possible scope of industrial policies.

As Donaldson explains, “An empirical lever here is to ask: If something makes a country’s sectors bigger, do they look more productive? If so, they would start exporting more to other countries. We reverse-engineer that.”

Costinot adds: “We are using that idea that if productivity is going up, that should be reflected in export patterns. The smoking gun for the existence of scale effects is that larger domestic markets go hand in hand with more exports.”

Ultimately, the scholars analyzed data for 61 countries at different points in time over the last few decades, with exports for 15 manufacturing sectors included. The figure of 1.08 percent long-run GDP gains is an average, with countries realizing gains ranging from 0.59 percent to 2.06 percent annually under favorable conditions. Smaller countries that are open to trade may realize larger proportional effects as well.

“We’re doing this global analysis and trying to be right on average,” Donaldson says. “It’s possible there are larger gains from industrial policy in particular settings.”

The study also suggests countries have greater room to redirect economic activity, based on varying levels of productivity among industries, than they can realistically enact due to relatively fixed demand. The paper estimates that if countries could fully reallocate workers to the industry with the largest room to grow, long-run welfare gains would be as high as 12.4 percent.

But that never happens. Suppose a country’s industrial policy helped one sector double in size while becoming 20 percent more productive. In theory, the government should continue to back that industry. In reality, growth would slow as markets became saturated.

“That would be a pretty big scale effect,” Donaldson says. “But notice that in doubling the size of an industry, many forces would push back. Maybe consumers don’t want to consume twice as many manufactured goods. Just because there are large spillovers in productivity doesn’t mean optimally designed industrial policy has huge effects. It has to be in a world where people want those goods.”

Place-based policy

Costinot and Donaldson both emphasize that this study does not address all the possible factors that can be weighed either in favor of industrial policy or against it. Some governments might favor industrial policy as a way of evening out wage distributions and wealth inequality, fixing other market failures such as environmental damages or furthering strategic geopolitical goals. In the U.S., industrial policy has sometimes been viewed as a way of revitalizing recently deindustrialized areas while reskilling workers.

In charting the limits on industrial policy stemming from fairly fixed demand, the study touches on still bigger issues concerning global demand and restrictions on growth of any kind. Without increasing demand, enterprise of all kinds encounters size limits.

The outcome of the paper, in any case, is not necessarily a final conclusion about industrial policy, but deeper insight into its dynamics. As the authors note, the findings leave open the possibility that targeted interventions in specific sectors and specific regions could be very beneficial, when policy and trade conditions are right. Policymakers should grasp the amount of growth likely to result, however.

As Costinot notes, “The conclusion is not that there is no potential gain from industrial policy, but just that the textbook case doesn’t seem to be there.” At least, not to the extent some have assumed.

The research was supported, in part, by the U.S. National Science Foundation.

de MIT News https://ift.tt/YlzSw2m

When MIT’s interdisciplinary NEET program is a perfect fit

At an early age, Katie Spivakovsky learned to study the world from different angles. Dinner-table conversations at her family’s home in Menlo Park, California, often leaned toward topics like the Maillard reaction — the chemistry behind food browning — or the fascinating mysteries of prime numbers. Spivakovsky’s parents, one of whom studied physical chemistry and the other statistics, fostered a love of knowledge that crossed disciplines. 

In high school, Spivakovsky explored it all, from classical literature to computer science. She knew she wanted an undergraduate experience that encouraged her broad interests, a place where every field was within reach. 

“MIT immediately stood out,” Spivakovsky says. “But it was specifically the existence of New Engineering Education Transformation (NEET) — a truly unique initiative that immerses undergraduates in interdisciplinary opportunities both within and beyond campus — that solidified my belief that MIT was the perfect fit for me.”  

NEET is a cross-departmental education program that empowers undergraduates to tackle the pressing challenges of the 21st century through interdisciplinary learning. Starting in their sophomore year, NEET scholars choose from one of four domains of study, or “threads:” Autonomous Machines, Climate and Sustainability Systems, Digital Cities, or Living Machines. After the typical four years, NEET scholars graduate with a degree in their major and a NEET certificate, equipping them with both depth in their chosen field and the ability to work in, and drive impact across, multiple domains. 

Spivakovsky is now a junior double-majoring in biological engineering and artificial intelligence and decision-making, with a minor in mathematics. At a time when fields like biology and computer science are merging like never before, she describes herself as “interested in leveraging engineering and computational tools to discover new biomedical insights” — a central theme of NEET’s Living Machines thread, in which she is now enrolled. 

“NEET is about more than engineering,” says Amitava “Babi” Mitra, NEET founding executive director. “It’s about nurturing young engineers who dream big, value collaboration, and are ready to tackle the world’s toughest challenges with heart and curiosity. Watching students like Katie thrive is why this program matters so deeply.”  

Spivakovsky’s achievements while at MIT already have a global reach. In 2023, she led an undergraduate team at the International Genetically Engineered Machine (iGEM) competition in Paris, France, where they presented a proof of concept for a therapy to treat cancer cachexia. Cachexia is a fat- and muscle-wasting condition with no FDA-approved treatment. The condition affects 80 percent of late-stage cancer patients and is responsible for 30 percent of cancer deaths. Spivakovsky’s team won a silver medal for proposing the engineering of macrophages to remove excess interleukin-6, a pro-inflammatory protein overproduced in cachexia patients, and their research was later published in MIT’s Undergraduate Research Journal, an honor she says was “unreal and humbling.”  

Spivakovsky works as a student researcher in the BioNanoLab of Mark Bathe, professor of biological engineering and former NEET faculty director. The lab uses DNA and RNA to engineer nanoscale materials for such uses as therapeutics and computing. Her focus is validating nucleic acid nanoparticles for use in therapeutics. 

According to Bathe, “Katie shows tremendous promise as a scientific leader — she brings unparalleled passion and creativity to her project on making novel vaccines with a depth of knowledge in both biology and computation that is truly unmatched.” 

Spivakovsky says class 20.054 (Living Machines Research Immersion), which she is taking in the NEET program, complements her work in Bathe’s lab and provides well-rounded experience through workshops that emphasize scientific communication, staying abreast of scientific literature, and research progress updates. “I’m interested in a range of subjects and find that switching between them helps keep things fresh,” she says.  

Her interdisciplinary drive took her to Merck over the summer, where Spivakovsky interned on the Modeling and Informatics team. While contributing to the development of a drug to deactivate a cancer-causing protein, she says she learned to use computational chemistry tools and developed geometric analysis techniques to identify locations on the protein where drug molecules might be able to bind.  

“My team continues to actively use the software I developed and the insights I gained through my work,” Spivakovsky says. “The target protein has an enormous patient population, so I am hopeful that within the next decade, drugs will enter the market, and my small contribution may make a difference in many lives.”  

As she looks toward her future, Spivakovsky envisions herself at the intersection of artificial intelligence and biology, ideally in a role that combines wet lab with computational research. “I can’t see myself in a career entirely devoid of one or the other,” she says. “This incredible synergy is where I feel most inspired.”   

Wherever Spivakovsky’s curiosity leads her next, she says one thing is certain: “NEET has really helped my development as a scientist.” 

de MIT News https://ift.tt/Yd6Bvwc

3 Questions: Tracking MIT graduates’ career trajectories

In a fall letter to MIT alumni, President Sally Kornbluth wrote: “[T]he world has never been more ready to reward our graduates for what they know — and know how to do.” During her tenure leading MIT Career Advising and Professional Development (CAPD), Deborah Liverman has seen firsthand how — and how well — MIT undergraduate and graduate students leverage their education to make an impact around the globe in academia, industry, entrepreneurship, medicine, government and nonprofits, and other professions. Here, Liverman shares her observations about trends in students’ career paths and the complexities of the job market they must navigate along the way.

Q: How do our students fare when they graduate from MIT?

A: We routinely survey our undergraduates and graduate students to track post-graduation outcomes, so fortunately we have a wealth of data. And ultimately, this enables us to stay on top of changes from year to year and to serve our students better.

The short answer is that our students fare exceptionally well when they leave the Institute! In our 2023 Graduating Student Survey, which is an exit survey for bachelor’s degree and master’s degree students, 49 percent of bachelor’s respondents and 79 percent of master’s respondents entered the workforce after graduating, and 43 percent and 14 percent started graduate school programs, respectively. Among those seeking immediate employment, 92 percent of bachelor’s and 87 percent of master’s degree students reported obtaining a job within three months of graduation.

What is notable, and frankly, wonderful, is that these two cohorts really took advantage of the rich ecosystem of experiential learning opportunities we have at MIT. The majority of Class of 2023 seniors participated in some form of experiential learning before graduation: 94 percent of them had a UROP [Undergraduate Research Opportunities Program], 75 percent interned, 66 percent taught or tutored, and 38 percent engaged with or mentored at campus makerspaces. Among master’s degree graduates in 2023, 56 percent interned, 45 percent taught or tutored, and 30 percent took part in entrepreneurial ventures or activities. About 47 percent of bachelor’s graduates said that a previous internship or externship led to the offer that they accepted, and 46 percent of master’s graduates are a founding member of a company.

We conduct a separate survey for doctoral students. I think there’s a common misperception that most of our PhD students go into academia. But a sizable portion choose not to stay in the academy. According to our 2024 Doctoral Exit Survey, 41 percent of graduates planned to go into industry. As of the survey date, of those who were going on to employment, 76 percent had signed a contract or made a definite commitment to a postdoc or other work, and only 9 percent were seeking a position but had no specific prospects.

A cohort of students, as well as some alumni, work with CAPD’s Prehealth Advising staff to apply for medical school. Last year we supported 73 students and alumni consisting of 25 undergrads, eight graduate students, and 40 alumni, with an acceptance rate of 79 percent — well above the national rate of 41 percent.

Q: How does CAPD work with students and postdocs to cultivate their professional development and help them evaluate their career options?

A: As you might expect, the career and graduate school landscape is constantly changing. In turn, CAPD strives to continuously evolve, so that we can best support and prepare our students. It certainly keeps us on our feet!

One of the things we have changed recently is our fundamental approach to working with students. We migrated our advising model from a major-specific focus to instead center on career interest areas. That allows us to prioritize skills and use a cross-disciplinary approach to advising students. So when an advisor sits down (or Zooms) with a student, that one-on-one session creates plenty of space to discuss a student’s individual values, goals, and other career-decision influencing factors.

I would say that another area we have been heavily focused on is providing new ways for students to explore careers. To that end, we developed two roles — an assistant director of career exploration and an assistant director of career prototype — to support new initiatives. And we provide career exploration fellowships and grants for undergraduate and graduate students so that they can explore fields that may be niche to MIT.

Career exploration is really important, but we want to meet students and postdocs where they are. We know they are incredibly busy at MIT, so our goal is to provide a variety of formats to make that possible, from a one-hour workshop or speaker, to a daylong shadowing experience, or a longer-term internship. For example, we partnered with departments to create the Career Exploration Series and the Infinite Careers speaker series, where we show students various avenues to get to a career. We have also created more opportunities to interact with alumni or other employers through one-day shadowing opportunities, micro-internships, internships, and employer coffee chats. The Prehealth Advising program I mentioned before offers many avenues to explore the field of medicine, so students can really make informed decisions about the path they want to pursue.

We are also looking at our existing programming to identify opportunities to build in career exploration, such as the Fall Career Fair. We have been working on identifying employers who are open to having career exploration conversations with — or hiring — first-year undergraduates, with access to these employers 30 minutes before the start of the fair. This year, the fair drew 4,400 candidates (students, postdocs, and alumni) and 180 employers, so it’s a great opportunity to leverage an event we already have in place and make it even more fruitful for both students and employers.

I do want to underscore that career exploration is just as important for graduate students as it is for undergraduates. In the doctoral exit survey I mentioned, 37 percent of 2024 graduates said they had changed their mind about the type of employer for whom they expected to work since entering their graduate program, and 38 percent had changed their mind about the type of position they expected to have. CAPD has developed exploration programming geared specifically for them, such as the CHAOS Process and our Graduate Student Professional Development offerings.

Q: What kinds of trends are you seeing in the current job market? And as students receive job offers, how do they weigh factors like the ethical considerations of working for a certain company or industry, the political landscape in the U.S. and abroad, the climate impact of a certain company or industry, or other issues?

A: Well, one notable trend is just the sheer volume of job applications. With platforms like LinkedIn’s Easy Apply, it’s easier for job seekers to apply to hundreds of jobs at once. Employers and organizations have more candidates, so applicants have to do more to stand out. Companies that, in the past, have had to seek out candidates are now deciding the best use of their recruiting efforts.

I would say the current job market is mixed. MIT students, graduates, and postdocs have experienced delayed job offers and starting dates pushed back in consulting and some tech firms. Companies are being intentional about recruiting and hiring college graduates. So students need to keep an open mind and not have their heart set on a particular employer. And if that employer isn’t hiring, then they may have to optimize their job search and consider other opportunities where they can gain experience.

On a more granular level, we do see trends in certain fields. Biotech has had a tough year, but there’s an uptick in opportunities in government, space, aerospace, and in the climate/sustainability and energy sectors. Companies are increasingly adopting AI in their business practices, so they’re hiring in that area. And financial services is a hot market for MIT candidates with strong technical skills.

As for how a student evaluates a job offer, according to the Graduating Student Survey, students look at many factors, including the job content, fit with the employer’s culture, opportunity for career advancement, and of course salary. However, students are also interested in exploring how an organization fits with their values.

CAPD provides various opportunities and resources to help them zero in on what matters most to them, from on-demand resources to one-on-one sessions with our advisors. As they research potential companies, we encourage them to make the most of career fairs and recruiting events. Throughout the academic year, MIT hosts and collaborates on over a dozen career fairs and large recruiting events. Companies are invited based on MIT candidates’ interests. The variety of opportunities means students can connect with different industries, explore careers, and apply to internships, jobs and research opportunities.

We also recommend that they take full advantage of MIT’s curated instance of Handshake, an online recruiting platform for higher education students and alumni. CAPD has collaborated with offices and groups to create filters and identifiers in Handshake to help candidates decide what is important to them, such as a company’s commitment to inclusive practices or their sustainability initiatives.

As advisors, we encourage each student to think about which factors are important for them when evaluating job offers and determine if an employer aligns with their values and goals. And we encourage and honor each student’s right to include those values and goals in their career decision-making process. Accepting a job is a very personal decision, and we are here to support each student every step of the way.

de MIT News https://ift.tt/Zdi1WRf