More stories

  • in

    Six MIT students selected as spring 2024 MIT-Pillar AI Collective Fellows

    The MIT-Pillar AI Collective has announced six fellows for the spring 2024 semester. With support from the program, the graduate students, who are in their final year of a master’s or PhD program, will conduct research in the areas of AI, machine learning, and data science with the aim of commercializing their innovations.

    Launched by MIT’s School of Engineering and Pillar VC in 2022, the MIT-Pillar AI Collective supports faculty, postdocs, and students conducting research on AI, machine learning, and data science. Supported by a gift from Pillar VC and administered by the MIT Deshpande Center for Technological Innovation, the mission of the program is to advance research toward commercialization.

    The spring 2024 MIT-Pillar AI Collective Fellows are:

    Yasmeen AlFaraj

    Yasmeen AlFaraj is a PhD candidate in chemistry whose interest is in the application of data science and machine learning to soft materials design to enable next-generation, sustainable plastics, rubber, and composite materials. More specifically, she is applying machine learning to the design of novel molecular additives to enable the low-cost manufacturing of chemically deconstructable thermosets and composites. AlFaraj’s work has led to the discovery of scalable, translatable new materials that could address thermoset plastic waste. As a Pillar Fellow, she will pursue bringing this technology to market, initially focusing on wind turbine blade manufacturing and conformal coatings. Through the Deshpande Center for Technological Innovation, AlFaraj serves as a lead for a team developing a spinout focused on recyclable versions of existing high-performance thermosets by incorporating small quantities of a degradable co-monomer. In addition, she participated in the National Science Foundation Innovation Corps program and recently graduated from the Clean Tech Open, where she focused on enhancing her business plan, analyzing potential markets, ensuring a complete IP portfolio, and connecting with potential funders. AlFaraj earned a BS in chemistry from University of California at Berkeley.

    Ruben Castro Ornelas

    Ruben Castro Ornelas is a PhD student in mechanical engineering who is passionate about the future of multipurpose robots and designing the hardware to use them with AI control solutions. Combining his expertise in programming, embedded systems, machine design, reinforcement learning, and AI, he designed a dexterous robotic hand capable of carrying out useful everyday tasks without sacrificing size, durability, complexity, or simulatability. Ornelas’s innovative design holds significant commercial potential in domestic, industrial, and health-care applications because it could be adapted to hold everything from kitchenware to delicate objects. As a Pillar Fellow, he will focus on identifying potential commercial markets, determining the optimal approach for business-to-business sales, and identifying critical advisors. Ornelas served as co-director of StartLabs, an undergraduate entrepreneurship club at MIT, where he earned an BS in mechanical engineering.

    Keeley Erhardt

    Keeley Erhardt is a PhD candidate in media arts and sciences whose research interests lie in the transformative potential of AI in network analysis, particularly for entity correlation and hidden link detection within and across domains. She has designed machine learning algorithms to identify and track temporal correlations and hidden signals in large-scale networks, uncovering online influence campaigns originating from multiple countries. She has similarly demonstrated the use of graph neural networks to identify coordinated cryptocurrency accounts by analyzing financial time series data and transaction dynamics. As a Pillar Fellow, Erhardt will pursue the potential commercial applications of her work, such as detecting fraud, propaganda, money laundering, and other covert activity in the finance, energy, and national security sectors. She has had internships at Google, Facebook, and Apple and held software engineering roles at multiple tech unicorns. Erhardt earned an MEng in electrical engineering and computer science and a BS in computer science, both from MIT.

    Vineet Jagadeesan Nair

    Vineet Jagadeesan Nair is a PhD candidate in mechanical engineering whose research focuses on modeling power grids and designing electricity markets to integrate renewables, batteries, and electric vehicles. He is broadly interested in developing computational tools to tackle climate change. As a Pillar Fellow, Nair will explore the application of machine learning and data science to power systems. Specifically, he will experiment with approaches to improve the accuracy of forecasting electricity demand and supply with high spatial-temporal resolution. In collaboration with Project Tapestry @ Google X, he is also working on fusing physics-informed machine learning with conventional numerical methods to increase the speed and accuracy of high-fidelity simulations. Nair’s work could help realize future grids with high penetrations of renewables and other clean, distributed energy resources. Outside academics, Nair is active in entrepreneurship, most recently helping to organize the 2023 MIT Global Startup Workshop in Greece. He earned an MS in computational science and engineering from MIT, an MPhil in energy technologies from Cambridge University as a Gates Scholar, and a BS in mechanical engineering and a BA in economics from University of California at Berkeley.

    Mahdi Ramadan

    Mahdi Ramadan is a PhD candidate in brain and cognitive sciences whose research interests lie at the intersection of cognitive science, computational modeling, and neural technologies. His work uses novel unsupervised methods for learning and generating interpretable representations of neural dynamics, capitalizing on recent advances in AI, specifically contrastive and geometric deep learning techniques capable of uncovering the latent dynamics underlying neural processes with high fidelity. As a Pillar Fellow, he will leverage these methods to gain a better understanding of dynamical models of muscle signals for generative motor control. By supplementing current spinal prosthetics with generative AI motor models that can streamline, speed up, and correct limb muscle activations in real time, as well as potentially using multimodal vision-language models to infer the patients’ high-level intentions, Ramadan aspires to build truly scalable, accessible, and capable commercial neuroprosthetics. Ramadan’s entrepreneurial experience includes being the co-founder of UltraNeuro, a neurotechnology startup, and co-founder of Presizely, a computer vision startup. He earned a BS in neurobiology from University of Washington.

    Rui (Raymond) Zhou

    Rui (Raymond) Zhou is a PhD candidate in mechanical engineering whose research focuses on multimodal AI for engineering design. As a Pillar Fellow, he will advance models that could enable designers to translate information in any modality or combination of modalities into comprehensive 2D and 3D designs, including parametric data, component visuals, assembly graphs, and sketches. These models could also optimize existing human designs to accomplish goals such as improving ergonomics or reducing drag coefficient. Ultimately, Zhou aims to translate his work into a software-as-a-service platform that redefines product design across various sectors, from automotive to consumer electronics. His efforts have the potential to not only accelerate the design process but also reduce costs, opening the door to unprecedented levels of customization, idea generation, and rapid prototyping. Beyond his academic pursuits, Zhou founded UrsaTech, a startup that integrates AI into education and engineering design. He earned a BS in electrical engineering and computer sciences from University of California at Berkeley. More

  • in

    How symmetry can come to the aid of machine learning

    Behrooz Tahmasebi — an MIT PhD student in the Department of Electrical Engineering and Computer Science (EECS) and an affiliate of the Computer Science and Artificial Intelligence Laboratory (CSAIL) — was taking a mathematics course on differential equations in late 2021 when a glimmer of inspiration struck. In that class, he learned for the first time about Weyl’s law, which had been formulated 110 years earlier by the German mathematician Hermann Weyl. Tahmasebi realized it might have some relevance to the computer science problem he was then wrestling with, even though the connection appeared — on the surface — to be thin, at best. Weyl’s law, he says, provides a formula that measures the complexity of the spectral information, or data, contained within the fundamental frequencies of a drum head or guitar string.

    Tahmasebi was, at the same time, thinking about measuring the complexity of the input data to a neural network, wondering whether that complexity could be reduced by taking into account some of the symmetries inherent to the dataset. Such a reduction, in turn, could facilitate — as well as speed up — machine learning processes.

    Weyl’s law, conceived about a century before the boom in machine learning, had traditionally been applied to very different physical situations — such as those concerning the vibrations of a string or the spectrum of electromagnetic (black-body) radiation given off by a heated object. Nevertheless, Tahmasebi believed that a customized version of that law might help with the machine learning problem he was pursuing. And if the approach panned out, the payoff could be considerable.

    He spoke with his advisor, Stefanie Jegelka — an associate professor in EECS and affiliate of CSAIL and the MIT Institute for Data, Systems, and Society — who believed the idea was definitely worth looking into. As Tahmasebi saw it, Weyl’s law had to do with gauging the complexity of data, and so did this project. But Weyl’s law, in its original form, said nothing about symmetry.

    He and Jegelka have now succeeded in modifying Weyl’s law so that symmetry can be factored into the assessment of a dataset’s complexity. “To the best of my knowledge,” Tahmasebi says, “this is the first time Weyl’s law has been used to determine how machine learning can be enhanced by symmetry.”

    The paper he and Jegelka wrote earned a “Spotlight” designation when it was presented at the December 2023 conference on Neural Information Processing Systems — widely regarded as the world’s top conference on machine learning.

    This work, comments Soledad Villar, an applied mathematician at Johns Hopkins University, “shows that models that satisfy the symmetries of the problem are not only correct but also can produce predictions with smaller errors, using a small amount of training points. [This] is especially important in scientific domains, like computational chemistry, where training data can be scarce.”

    In their paper, Tahmasebi and Jegelka explored the ways in which symmetries, or so-called “invariances,” could benefit machine learning. Suppose, for example, the goal of a particular computer run is to pick out every image that contains the numeral 3. That task can be a lot easier, and go a lot quicker, if the algorithm can identify the 3 regardless of where it is placed in the box — whether it’s exactly in the center or off to the side — and whether it is pointed right-side up, upside down, or oriented at a random angle. An algorithm equipped with the latter capability can take advantage of the symmetries of translation and rotations, meaning that a 3, or any other object, is not changed in itself by altering its position or by rotating it around an arbitrary axis. It is said to be invariant to those shifts. The same logic can be applied to algorithms charged with identifying dogs or cats. A dog is a dog is a dog, one might say, irrespective of how it is embedded within an image. 

    The point of the entire exercise, the authors explain, is to exploit a dataset’s intrinsic symmetries in order to reduce the complexity of machine learning tasks. That, in turn, can lead to a reduction in the amount of data needed for learning. Concretely, the new work answers the question: How many fewer data are needed to train a machine learning model if the data contain symmetries?

    There are two ways of achieving a gain, or benefit, by capitalizing on the symmetries present. The first has to do with the size of the sample to be looked at. Let’s imagine that you are charged, for instance, with analyzing an image that has mirror symmetry — the right side being an exact replica, or mirror image, of the left. In that case, you don’t have to look at every pixel; you can get all the information you need from half of the image — a factor of two improvement. If, on the other hand, the image can be partitioned into 10 identical parts, you can get a factor of 10 improvement. This kind of boosting effect is linear.

    To take another example, imagine you are sifting through a dataset, trying to find sequences of blocks that have seven different colors — black, blue, green, purple, red, white, and yellow. Your job becomes much easier if you don’t care about the order in which the blocks are arranged. If the order mattered, there would be 5,040 different combinations to look for. But if all you care about are sequences of blocks in which all seven colors appear, then you have reduced the number of things — or sequences — you are searching for from 5,040 to just one.

    Tahmasebi and Jegelka discovered that it is possible to achieve a different kind of gain — one that is exponential — that can be reaped for symmetries that operate over many dimensions. This advantage is related to the notion that the complexity of a learning task grows exponentially with the dimensionality of the data space. Making use of a multidimensional symmetry can therefore yield a disproportionately large return. “This is a new contribution that is basically telling us that symmetries of higher dimension are more important because they can give us an exponential gain,” Tahmasebi says. 

    The NeurIPS 2023 paper that he wrote with Jegelka contains two theorems that were proved mathematically. “The first theorem shows that an improvement in sample complexity is achievable with the general algorithm we provide,” Tahmasebi says. The second theorem complements the first, he added, “showing that this is the best possible gain you can get; nothing else is achievable.”

    He and Jegelka have provided a formula that predicts the gain one can obtain from a particular symmetry in a given application. A virtue of this formula is its generality, Tahmasebi notes. “It works for any symmetry and any input space.” It works not only for symmetries that are known today, but it could also be applied in the future to symmetries that are yet to be discovered. The latter prospect is not too farfetched to consider, given that the search for new symmetries has long been a major thrust in physics. That suggests that, as more symmetries are found, the methodology introduced by Tahmasebi and Jegelka should only get better over time.

    According to Haggai Maron, a computer scientist at Technion (the Israel Institute of Technology) and NVIDIA who was not involved in the work, the approach presented in the paper “diverges substantially from related previous works, adopting a geometric perspective and employing tools from differential geometry. This theoretical contribution lends mathematical support to the emerging subfield of ‘Geometric Deep Learning,’ which has applications in graph learning, 3D data, and more. The paper helps establish a theoretical basis to guide further developments in this rapidly expanding research area.” More

  • in

    Creating new skills and new connections with MIT’s Quantitative Methods Workshop

    Starting on New Year’s Day, when many people were still clinging to holiday revelry, scores of students and faculty members from about a dozen partner universities instead flipped open their laptops for MIT’s Quantitative Methods Workshop, a jam-packed, weeklong introduction to how computational and mathematical techniques can be applied to neuroscience and biology research. But don’t think of QMW as a “crash course.” Instead the program’s purpose is to help elevate each participant’s scientific outlook, both through the skills and concepts it imparts and the community it creates.

    “It broadens their horizons, it shows them significant applications they’ve never thought of, and introduces them to people whom as researchers they will come to know and perhaps collaborate with one day,” says Susan L. Epstein, a Hunter College computer science professor and education coordinator of MIT’s Center for Brains, Minds, and Machines, which hosts the program with the departments of Biology and Brain and Cognitive Sciences and The Picower Institute for Learning and Memory. “It is a model of interdisciplinary scholarship.”

    This year 83 undergraduates and faculty members from institutions that primarily serve groups underrepresented in STEM fields took part in the QMW, says organizer Mandana Sassanfar, senior lecturer and director of diversity and science outreach across the four hosting MIT entities. Since the workshop launched in 2010, it has engaged more than 1,000 participants, of whom more than 170 have gone on to participate in MIT Summer Research Programs (such as MSRP-BIO), and 39 have come to MIT for graduate school.

    Individual goals, shared experience

    Undergraduates and faculty in various STEM disciplines often come to QMW to gain an understanding of, or expand their expertise in, computational and mathematical data analysis. Computer science- and statistics-minded participants come to learn more about how such techniques can be applied in life sciences fields. In lectures; in hands-on labs where they used the computer programming language Python to process, analyze, and visualize data; and in less formal settings such as tours and lunches with MIT faculty, participants worked and learned together, and informed each other’s perspectives.

    Brain and Cognitive Sciences Professor Nancy Kanwisher delivers a lecture in MIT’s Building 46 on functional brain imaging to QMW participants.

    Photo: Mandana Sassanfar

    Previous item
    Next item

    And regardless of their field of study, participants made connections with each other and with the MIT students and faculty who taught and spoke over the course of the week.

    Hunter College computer science sophomore Vlad Vostrikov says that while he has already worked with machine learning and other programming concepts, he was interested to “branch out” by seeing how they are used to analyze scientific datasets. He also valued the chance to learn the experiences of the graduate students who teach QMW’s hands-on labs.

    “This was a good way to explore computational biology and neuroscience,” Vostrikov says. “I also really enjoy hearing from the people who teach us. It’s interesting to hear where they come from and what they are doing.”

    Jariatu Kargbo, a biology and chemistry sophomore at University of Maryland Baltimore County, says when she first learned of the QMW she wasn’t sure it was for her. It seemed very computation-focused. But her advisor Holly Willoughby encouraged Kargbo to attend to learn about how programming could be useful in future research — currently she is taking part in research on the retina at UMBC. More than that, Kargbo also realized it would be a good opportunity to make connections at MIT in advance of perhaps applying for MSRP this summer.

    “I thought this would be a great way to meet up with faculty and see what the environment is like here because I’ve never been to MIT before,” Kargbo says. “It’s always good to meet other people in your field and grow your network.”

    QMW is not just for students. It’s also for their professors, who said they can gain valuable professional education for their research and teaching.

    Fayuan Wen, an assistant professor of biology at Howard University, is no stranger to computational biology, having performed big data genetic analyses of sickle cell disease (SCD). But she’s mostly worked with the R programming language and QMW’s focus is on Python. As she looks ahead to projects in which she wants analyze genomic data to help predict disease outcomes in SCD and HIV, she says a QMW session delivered by biology graduate student Hannah Jacobs was perfectly on point.

    “This workshop has the skills I want to have,” Wen says.

    Moreover, Wen says she is looking to start a machine-learning class in the Howard biology department and was inspired by some of the teaching materials she encountered at QMW — for example, online curriculum modules developed by Taylor Baum, an MIT graduate student in electrical engineering and computer science and Picower Institute labs, and Paloma Sánchez-Jáuregui, a coordinator who works with Sassanfar.

    Tiziana Ligorio, a Hunter College computer science doctoral lecturer who together with Epstein teaches a deep machine-learning class at the City University of New York campus, felt similarly. Rather than require a bunch of prerequisites that might drive students away from the class, Ligorio was looking to QMW’s intense but introductory curriculum as a resource for designing a more inclusive way of getting students ready for the class.

    Instructive interactions

    Each day runs from 9 a.m. to 5 p.m., including morning and afternoon lectures and hands-on sessions. Class topics ranged from statistical data analysis and machine learning to brain-computer interfaces, brain imaging, signal processing of neural activity data, and cryogenic electron microscopy.

    “This workshop could not happen without dedicated instructors — grad students, postdocs, and faculty — who volunteer to give lectures, design and teach hands-on computer labs, and meet with students during the very first week of January,” Saassanfar says.

    MIT assistant professor of biology Brady Weissbourd (center) converses with QMW student participants during a lunch break.

    Photo: Mandana Sassanfar

    Previous item
    Next item

    The sessions surround student lunches with MIT faculty members. For example, at midday Jan. 2, assistant professor of biology Brady Weissbourd, an investigator in the Picower Institute, sat down with seven students in one of Building 46’s curved sofas to field questions about his neuroscience research in jellyfish and how he uses quantitative techniques as part of that work. He also described what it’s like to be a professor, and other topics that came to the students’ minds.

    Then the participants all crossed Vassar Street to Building 26’s Room 152, where they formed different but similarly sized groups for the hands-on lab “Machine learning applications to studying the brain,” taught by Baum. She guided the class through Python exercises she developed illustrating “supervised” and “unsupervised” forms of machine learning, including how the latter method can be used to discern what a person is seeing based on magnetic readings of brain activity.

    As students worked through the exercises, tablemates helped each other by supplementing Baum’s instruction. Ligorio, Vostrikov, and Kayla Blincow, assistant professor of biology at the University of the Virgin Islands, for instance, all leapt to their feet to help at their tables.

    Hunter College lecturer of computer science Tiziana Ligorio (standing) explains a Python programming concept to students at her table during a workshop session.

    Photo: David Orenstein

    Previous item
    Next item

    At the end of the class, when Baum asked students what they had learned, they offered a litany of new knowledge. Survey data that Sassanfar and Sánchez-Jáuregui use to anonymously track QMW outcomes, revealed many more such attestations of the value of the sessions. With a prompt asking how one might apply what they’ve learned, one respondent wrote: “Pursue a research career or endeavor in which I apply the concepts of computer science and neuroscience together.”

    Enduring connections

    While some new QMW attendees might only be able to speculate about how they’ll apply their new skills and relationships, Luis Miguel de Jesús Astacio could testify to how attending QMW as an undergraduate back in 2014 figured into a career where he is now a faculty member in physics at the University of Puerto Rico Rio Piedras Campus. After QMW, he returned to MIT that summer as a student in the lab of neuroscientist and Picower Professor Susumu Tonegawa. He came back again in 2016 to the lab of physicist and Francis Friedman Professor Mehran Kardar. What’s endured for the decade has been his connection to Sassanfar. So while he was once a student at QMW, this year he was back with a cohort of undergraduates as a faculty member.

    Michael Aldarondo-Jeffries, director of academic advancement programs at the University of Central Florida, seconded the value of the networking that takes place at QMW. He has brought students for a decade, including four this year. What he’s observed is that as students come together in settings like QMW or UCF’s McNair program, which helps to prepare students for graduate school, they become inspired about a potential future as researchers.

    “The thing that stands out is just the community that’s formed,” he says. “For many of the students, it’s the first time that they’re in a group that understands what they’re moving toward. They don’t have to explain why they’re excited to read papers on a Friday night.”

    Or why they are excited to spend a week including New Year’s Day at MIT learning how to apply quantitative methods to life sciences data. More

  • in

    New hope for early pancreatic cancer intervention via AI-based risk prediction

    The first documented case of pancreatic cancer dates back to the 18th century. Since then, researchers have undertaken a protracted and challenging odyssey to understand the elusive and deadly disease. To date, there is no better cancer treatment than early intervention. Unfortunately, the pancreas, nestled deep within the abdomen, is particularly elusive for early detection. 

    MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) scientists, alongside Limor Appelbaum, a staff scientist in the Department of Radiation Oncology at Beth Israel Deaconess Medical Center (BIDMC), were eager to better identify potential high-risk patients. They set out to develop two machine-learning models for early detection of pancreatic ductal adenocarcinoma (PDAC), the most common form of the cancer. To access a broad and diverse database, the team synced up with a federated network company, using electronic health record data from various institutions across the United States. This vast pool of data helped ensure the models’ reliability and generalizability, making them applicable across a wide range of populations, geographical locations, and demographic groups.

    The two models — the “PRISM” neural network, and the logistic regression model (a statistical technique for probability), outperformed current methods. The team’s comparison showed that while standard screening criteria identify about 10 percent of PDAC cases using a five-times higher relative risk threshold, Prism can detect 35 percent of PDAC cases at this same threshold. 

    Using AI to detect cancer risk is not a new phenomena — algorithms analyze mammograms, CT scans for lung cancer, and assist in the analysis of Pap smear tests and HPV testing, to name a few applications. “The PRISM models stand out for their development and validation on an extensive database of over 5 million patients, surpassing the scale of most prior research in the field,” says Kai Jia, an MIT PhD student in electrical engineering and computer science (EECS), MIT CSAIL affiliate, and first author on an open-access paper in eBioMedicine outlining the new work. “The model uses routine clinical and lab data to make its predictions, and the diversity of the U.S. population is a significant advancement over other PDAC models, which are usually confined to specific geographic regions, like a few health-care centers in the U.S. Additionally, using a unique regularization technique in the training process enhanced the models’ generalizability and interpretability.” 

    “This report outlines a powerful approach to use big data and artificial intelligence algorithms to refine our approach to identifying risk profiles for cancer,” says David Avigan, a Harvard Medical School professor and the cancer center director and chief of hematology and hematologic malignancies at BIDMC, who was not involved in the study. “This approach may lead to novel strategies to identify patients with high risk for malignancy that may benefit from focused screening with the potential for early intervention.” 

    Prismatic perspectives

    The journey toward the development of PRISM began over six years ago, fueled by firsthand experiences with the limitations of current diagnostic practices. “Approximately 80-85 percent of pancreatic cancer patients are diagnosed at advanced stages, where cure is no longer an option,” says senior author Appelbaum, who is also a Harvard Medical School instructor as well as radiation oncologist. “This clinical frustration sparked the idea to delve into the wealth of data available in electronic health records (EHRs).”The CSAIL group’s close collaboration with Appelbaum made it possible to understand the combined medical and machine learning aspects of the problem better, eventually leading to a much more accurate and transparent model. “The hypothesis was that these records contained hidden clues — subtle signs and symptoms that could act as early warning signals of pancreatic cancer,” she adds. “This guided our use of federated EHR networks in developing these models, for a scalable approach for deploying risk prediction tools in health care.”Both PrismNN and PrismLR models analyze EHR data, including patient demographics, diagnoses, medications, and lab results, to assess PDAC risk. PrismNN uses artificial neural networks to detect intricate patterns in data features like age, medical history, and lab results, yielding a risk score for PDAC likelihood. PrismLR uses logistic regression for a simpler analysis, generating a probability score of PDAC based on these features. Together, the models offer a thorough evaluation of different approaches in predicting PDAC risk from the same EHR data.

    One paramount point for gaining the trust of physicians, the team notes, is better understanding how the models work, known in the field as interpretability. The scientists pointed out that while logistic regression models are inherently easier to interpret, recent advancements have made deep neural networks somewhat more transparent. This helped the team to refine the thousands of potentially predictive features derived from EHR of a single patient to approximately 85 critical indicators. These indicators, which include patient age, diabetes diagnosis, and an increased frequency of visits to physicians, are automatically discovered by the model but match physicians’ understanding of risk factors associated with pancreatic cancer. 

    The path forward

    Despite the promise of the PRISM models, as with all research, some parts are still a work in progress. U.S. data alone are the current diet for the models, necessitating testing and adaptation for global use. The path forward, the team notes, includes expanding the model’s applicability to international datasets and integrating additional biomarkers for more refined risk assessment.

    “A subsequent aim for us is to facilitate the models’ implementation in routine health care settings. The vision is to have these models function seamlessly in the background of health care systems, automatically analyzing patient data and alerting physicians to high-risk cases without adding to their workload,” says Jia. “A machine-learning model integrated with the EHR system could empower physicians with early alerts for high-risk patients, potentially enabling interventions well before symptoms manifest. We are eager to deploy our techniques in the real world to help all individuals enjoy longer, healthier lives.” 

    Jia wrote the paper alongside Applebaum and MIT EECS Professor and CSAIL Principal Investigator Martin Rinard, who are both senior authors of the paper. Researchers on the paper were supported during their time at MIT CSAIL, in part, by the Defense Advanced Research Projects Agency, Boeing, the National Science Foundation, and Aarno Labs. TriNetX provided resources for the project, and the Prevent Cancer Foundation also supported the team. More

  • in

    Multiple AI models help robots execute complex plans more transparently

    Your daily to-do list is likely pretty straightforward: wash the dishes, buy groceries, and other minutiae. It’s unlikely you wrote out “pick up the first dirty dish,” or “wash that plate with a sponge,” because each of these miniature steps within the chore feels intuitive. While we can routinely complete each step without much thought, a robot requires a complex plan that involves more detailed outlines.

    MIT’s Improbable AI Lab, a group within the Computer Science and Artificial Intelligence Laboratory (CSAIL), has offered these machines a helping hand with a new multimodal framework: Compositional Foundation Models for Hierarchical Planning (HiP), which develops detailed, feasible plans with the expertise of three different foundation models. Like OpenAI’s GPT-4, the foundation model that ChatGPT and Bing Chat were built upon, these foundation models are trained on massive quantities of data for applications like generating images, translating text, and robotics.Unlike RT2 and other multimodal models that are trained on paired vision, language, and action data, HiP uses three different foundation models each trained on different data modalities. Each foundation model captures a different part of the decision-making process and then works together when it’s time to make decisions. HiP removes the need for access to paired vision, language, and action data, which is difficult to obtain. HiP also makes the reasoning process more transparent.

    What’s considered a daily chore for a human can be a robot’s “long-horizon goal” — an overarching objective that involves completing many smaller steps first — requiring sufficient data to plan, understand, and execute objectives. While computer vision researchers have attempted to build monolithic foundation models for this problem, pairing language, visual, and action data is expensive. Instead, HiP represents a different, multimodal recipe: a trio that cheaply incorporates linguistic, physical, and environmental intelligence into a robot.

    “Foundation models do not have to be monolithic,” says NVIDIA AI researcher Jim Fan, who was not involved in the paper. “This work decomposes the complex task of embodied agent planning into three constituent models: a language reasoner, a visual world model, and an action planner. It makes a difficult decision-making problem more tractable and transparent.”The team believes that their system could help these machines accomplish household chores, such as putting away a book or placing a bowl in the dishwasher. Additionally, HiP could assist with multistep construction and manufacturing tasks, like stacking and placing different materials in specific sequences.Evaluating HiP

    The CSAIL team tested HiP’s acuity on three manipulation tasks, outperforming comparable frameworks. The system reasoned by developing intelligent plans that adapt to new information.

    First, the researchers requested that it stack different-colored blocks on each other and then place others nearby. The catch: Some of the correct colors weren’t present, so the robot had to place white blocks in a color bowl to paint them. HiP often adjusted to these changes accurately, especially compared to state-of-the-art task planning systems like Transformer BC and Action Diffuser, by adjusting its plans to stack and place each square as needed.

    Another test: arranging objects such as candy and a hammer in a brown box while ignoring other items. Some of the objects it needed to move were dirty, so HiP adjusted its plans to place them in a cleaning box, and then into the brown container. In a third demonstration, the bot was able to ignore unnecessary objects to complete kitchen sub-goals such as opening a microwave, clearing a kettle out of the way, and turning on a light. Some of the prompted steps had already been completed, so the robot adapted by skipping those directions.

    A three-pronged hierarchy

    HiP’s three-pronged planning process operates as a hierarchy, with the ability to pre-train each of its components on different sets of data, including information outside of robotics. At the bottom of that order is a large language model (LLM), which starts to ideate by capturing all the symbolic information needed and developing an abstract task plan. Applying the common sense knowledge it finds on the internet, the model breaks its objective into sub-goals. For example, “making a cup of tea” turns into “filling a pot with water,” “boiling the pot,” and the subsequent actions required.

    “All we want to do is take existing pre-trained models and have them successfully interface with each other,” says Anurag Ajay, a PhD student in the MIT Department of Electrical Engineering and Computer Science (EECS) and a CSAIL affiliate. “Instead of pushing for one model to do everything, we combine multiple ones that leverage different modalities of internet data. When used in tandem, they help with robotic decision-making and can potentially aid with tasks in homes, factories, and construction sites.”

    These models also need some form of “eyes” to understand the environment they’re operating in and correctly execute each sub-goal. The team used a large video diffusion model to augment the initial planning completed by the LLM, which collects geometric and physical information about the world from footage on the internet. In turn, the video model generates an observation trajectory plan, refining the LLM’s outline to incorporate new physical knowledge.This process, known as iterative refinement, allows HiP to reason about its ideas, taking in feedback at each stage to generate a more practical outline. The flow of feedback is similar to writing an article, where an author may send their draft to an editor, and with those revisions incorporated in, the publisher reviews for any last changes and finalizes.

    In this case, the top of the hierarchy is an egocentric action model, or a sequence of first-person images that infer which actions should take place based on its surroundings. During this stage, the observation plan from the video model is mapped over the space visible to the robot, helping the machine decide how to execute each task within the long-horizon goal. If a robot uses HiP to make tea, this means it will have mapped out exactly where the pot, sink, and other key visual elements are, and begin completing each sub-goal.Still, the multimodal work is limited by the lack of high-quality video foundation models. Once available, they could interface with HiP’s small-scale video models to further enhance visual sequence prediction and robot action generation. A higher-quality version would also reduce the current data requirements of the video models.That being said, the CSAIL team’s approach only used a tiny bit of data overall. Moreover, HiP was cheap to train and demonstrated the potential of using readily available foundation models to complete long-horizon tasks. “What Anurag has demonstrated is proof-of-concept of how we can take models trained on separate tasks and data modalities and combine them into models for robotic planning. In the future, HiP could be augmented with pre-trained models that can process touch and sound to make better plans,” says senior author Pulkit Agrawal, MIT assistant professor in EECS and director of the Improbable AI Lab. The group is also considering applying HiP to solving real-world long-horizon tasks in robotics.Ajay and Agrawal are lead authors on a paper describing the work. They are joined by MIT professors and CSAIL principal investigators Tommi Jaakkola, Joshua Tenenbaum, and Leslie Pack Kaelbling; CSAIL research affiliate and MIT-IBM AI Lab research manager Akash Srivastava; graduate students Seungwook Han and Yilun Du ’19; former postdoc Abhishek Gupta, who is now assistant professor at University of Washington; and former graduate student Shuang Li PhD ’23.

    The team’s work was supported, in part, by the National Science Foundation, the U.S. Defense Advanced Research Projects Agency, the U.S. Army Research Office, the U.S. Office of Naval Research Multidisciplinary University Research Initiatives, and the MIT-IBM Watson AI Lab. Their findings were presented at the 2023 Conference on Neural Information Processing Systems (NeurIPS). More

  • in

    Technique could efficiently solve partial differential equations for numerous applications

    In fields such as physics and engineering, partial differential equations (PDEs) are used to model complex physical processes to generate insight into how some of the most complicated physical and natural systems in the world function.

    To solve these difficult equations, researchers use high-fidelity numerical solvers, which can be very time-consuming and computationally expensive to run. The current simplified alternative, data-driven surrogate models, compute the goal property of a solution to PDEs rather than the whole solution. Those are trained on a set of data that has been generated by the high-fidelity solver, to predict the output of the PDEs for new inputs. This is data-intensive and expensive because complex physical systems require a large number of simulations to generate enough data. 

    In a new paper, “Physics-enhanced deep surrogates for partial differential equations,” published in December in Nature Machine Intelligence, a new method is proposed for developing data-driven surrogate models for complex physical systems in such fields as mechanics, optics, thermal transport, fluid dynamics, physical chemistry, and climate models.

    The paper was authored by MIT’s professor of applied mathematics Steven G. Johnson along with Payel Das and Youssef Mroueh of the MIT-IBM Watson AI Lab and IBM Research; Chris Rackauckas of Julia Lab; and Raphaël Pestourie, a former MIT postdoc who is now at Georgia Tech. The authors call their method “physics-enhanced deep surrogate” (PEDS), which combines a low-fidelity, explainable physics simulator with a neural network generator. The neural network generator is trained end-to-end to match the output of the high-fidelity numerical solver.

    “My aspiration is to replace the inefficient process of trial and error with systematic, computer-aided simulation and optimization,” says Pestourie. “Recent breakthroughs in AI like the large language model of ChatGPT rely on hundreds of billions of parameters and require vast amounts of resources to train and evaluate. In contrast, PEDS is affordable to all because it is incredibly efficient in computing resources and has a very low barrier in terms of infrastructure needed to use it.”

    In the article, they show that PEDS surrogates can be up to three times more accurate than an ensemble of feedforward neural networks with limited data (approximately 1,000 training points), and reduce the training data needed by at least a factor of 100 to achieve a target error of 5 percent. Developed using the MIT-designed Julia programming language, this scientific machine-learning method is thus efficient in both computing and data.

    The authors also report that PEDS provides a general, data-driven strategy to bridge the gap between a vast array of simplified physical models with corresponding brute-force numerical solvers modeling complex systems. This technique offers accuracy, speed, data efficiency, and physical insights into the process.

    Says Pestourie, “Since the 2000s, as computing capabilities improved, the trend of scientific models has been to increase the number of parameters to fit the data better, sometimes at the cost of a lower predictive accuracy. PEDS does the opposite by choosing its parameters smartly. It leverages the technology of automatic differentiation to train a neural network that makes a model with few parameters accurate.”

    “The main challenge that prevents surrogate models from being used more widely in engineering is the curse of dimensionality — the fact that the needed data to train a model increases exponentially with the number of model variables,” says Pestourie. “PEDS reduces this curse by incorporating information from the data and from the field knowledge in the form of a low-fidelity model solver.”

    The researchers say that PEDS has the potential to revive a whole body of the pre-2000 literature dedicated to minimal models — intuitive models that PEDS could make more accurate while also being predictive for surrogate model applications.

    “The application of the PEDS framework is beyond what we showed in this study,” says Das. “Complex physical systems governed by PDEs are ubiquitous, from climate modeling to seismic modeling and beyond. Our physics-inspired fast and explainable surrogate models will be of great use in those applications, and play a complementary role to other emerging techniques, like foundation models.”

    The research was supported by the MIT-IBM Watson AI Lab and the U.S. Army Research Office through the Institute for Soldier Nanotechnologies.  More

  • in

    Leveraging language to understand machines

    Natural language conveys ideas, actions, information, and intent through context and syntax; further, there are volumes of it contained in databases. This makes it an excellent source of data to train machine-learning systems on. Two master’s of engineering students in the 6A MEng Thesis Program at MIT, Irene Terpstra ’23 and Rujul Gandhi ’22, are working with mentors in the MIT-IBM Watson AI Lab to use this power of natural language to build AI systems.

    As computing is becoming more advanced, researchers are looking to improve the hardware that they run on; this means innovating to create new computer chips. And, since there is literature already available on modifications that can be made to achieve certain parameters and performance, Terpstra and her mentors and advisors Anantha Chandrakasan, MIT School of Engineering dean and the Vannevar Bush Professor of Electrical Engineering and Computer Science, and IBM’s researcher Xin Zhang, are developing an AI algorithm that assists in chip design.

    “I’m creating a workflow to systematically analyze how these language models can help the circuit design process. What reasoning powers do they have, and how can it be integrated into the chip design process?” says Terpstra. “And then on the other side, if that proves to be useful enough, [we’ll] see if they can automatically design the chips themselves, attaching it to a reinforcement learning algorithm.”

    To do this, Terpstra’s team is creating an AI system that can iterate on different designs. It means experimenting with various pre-trained large language models (like ChatGPT, Llama 2, and Bard), using an open-source circuit simulator language called NGspice, which has the parameters of the chip in code form, and a reinforcement learning algorithm. With text prompts, researchers will be able to query how the physical chip should be modified to achieve a certain goal in the language model and produced guidance for adjustments. This is then transferred into a reinforcement learning algorithm that updates the circuit design and outputs new physical parameters of the chip.

    “The final goal would be to combine the reasoning powers and the knowledge base that is baked into these large language models and combine that with the optimization power of the reinforcement learning algorithms and have that design the chip itself,” says Terpstra.

    Rujul Gandhi works with the raw language itself. As an undergraduate at MIT, Gandhi explored linguistics and computer sciences, putting them together in her MEng work. “I’ve been interested in communication, both between just humans and between humans and computers,” Gandhi says.

    Robots or other interactive AI systems are one area where communication needs to be understood by both humans and machines. Researchers often write instructions for robots using formal logic. This helps ensure that commands are being followed safely and as intended, but formal logic can be difficult for users to understand, while natural language comes easily. To ensure this smooth communication, Gandhi and her advisors Yang Zhang of IBM and MIT assistant professor Chuchu Fan are building a parser that converts natural language instructions into a machine-friendly form. Leveraging the linguistic structure encoded by the pre-trained encoder-decoder model T5, and a dataset of annotated, basic English commands for performing certain tasks, Gandhi’s system identifies the smallest logical units, or atomic propositions, which are present in a given instruction.

    “Once you’ve given your instruction, the model identifies all the smaller sub-tasks you want it to carry out,” Gandhi says. “Then, using a large language model, each sub-task can be compared against the available actions and objects in the robot’s world, and if any sub-task can’t be carried out because a certain object is not recognized, or an action is not possible, the system can stop right there to ask the user for help.”

    This approach of breaking instructions into sub-tasks also allows her system to understand logical dependencies expressed in English, like, “do task X until event Y happens.” Gandhi uses a dataset of step-by-step instructions across robot task domains like navigation and manipulation, with a focus on household tasks. Using data that are written just the way humans would talk to each other has many advantages, she says, because it means a user can be more flexible about how they phrase their instructions.

    Another of Gandhi’s projects involves developing speech models. In the context of speech recognition, some languages are considered “low resource” since they might not have a lot of transcribed speech available, or might not have a written form at all. “One of the reasons I applied to this internship at the MIT-IBM Watson AI Lab was an interest in language processing for low-resource languages,” she says. “A lot of language models today are very data-driven, and when it’s not that easy to acquire all of that data, that’s when you need to use the limited data efficiently.” 

    Speech is just a stream of sound waves, but humans having a conversation can easily figure out where words and thoughts start and end. In speech processing, both humans and language models use their existing vocabulary to recognize word boundaries and understand the meaning. In low- or no-resource languages, a written vocabulary might not exist at all, so researchers can’t provide one to the model. Instead, the model can make note of what sound sequences occur together more frequently than others, and infer that those might be individual words or concepts. In Gandhi’s research group, these inferred words are then collected into a pseudo-vocabulary that serves as a labeling method for the low-resource language, creating labeled data for further applications.

    The applications for language technology are “pretty much everywhere,” Gandhi says. “You could imagine people being able to interact with software and devices in their native language, their native dialect. You could imagine improving all the voice assistants that we use. You could imagine it being used for translation or interpretation.” More

  • in

    Image recognition accuracy: An unseen challenge confounding today’s AI

    Imagine you are scrolling through the photos on your phone and you come across an image that at first you can’t recognize. It looks like maybe something fuzzy on the couch; could it be a pillow or a coat? After a couple of seconds it clicks — of course! That ball of fluff is your friend’s cat, Mocha. While some of your photos could be understood in an instant, why was this cat photo much more difficult?

    MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) researchers were surprised to find that despite the critical importance of understanding visual data in pivotal areas ranging from health care to transportation to household devices, the notion of an image’s recognition difficulty for humans has been almost entirely ignored. One of the major drivers of progress in deep learning-based AI has been datasets, yet we know little about how data drives progress in large-scale deep learning beyond that bigger is better.

    In real-world applications that require understanding visual data, humans outperform object recognition models despite the fact that models perform well on current datasets, including those explicitly designed to challenge machines with debiased images or distribution shifts. This problem persists, in part, because we have no guidance on the absolute difficulty of an image or dataset. Without controlling for the difficulty of images used for evaluation, it’s hard to objectively assess progress toward human-level performance, to cover the range of human abilities, and to increase the challenge posed by a dataset.

    To fill in this knowledge gap, David Mayo, an MIT PhD student in electrical engineering and computer science and a CSAIL affiliate, delved into the deep world of image datasets, exploring why certain images are more difficult for humans and machines to recognize than others. “Some images inherently take longer to recognize, and it’s essential to understand the brain’s activity during this process and its relation to machine learning models. Perhaps there are complex neural circuits or unique mechanisms missing in our current models, visible only when tested with challenging visual stimuli. This exploration is crucial for comprehending and enhancing machine vision models,” says Mayo, a lead author of a new paper on the work.

    This led to the development of a new metric, the “minimum viewing time” (MVT), which quantifies the difficulty of recognizing an image based on how long a person needs to view it before making a correct identification. Using a subset of ImageNet, a popular dataset in machine learning, and ObjectNet, a dataset designed to test object recognition robustness, the team showed images to participants for varying durations from as short as 17 milliseconds to as long as 10 seconds, and asked them to choose the correct object from a set of 50 options. After over 200,000 image presentation trials, the team found that existing test sets, including ObjectNet, appeared skewed toward easier, shorter MVT images, with the vast majority of benchmark performance derived from images that are easy for humans.

    The project identified interesting trends in model performance — particularly in relation to scaling. Larger models showed considerable improvement on simpler images but made less progress on more challenging images. The CLIP models, which incorporate both language and vision, stood out as they moved in the direction of more human-like recognition.

    “Traditionally, object recognition datasets have been skewed towards less-complex images, a practice that has led to an inflation in model performance metrics, not truly reflective of a model’s robustness or its ability to tackle complex visual tasks. Our research reveals that harder images pose a more acute challenge, causing a distribution shift that is often not accounted for in standard evaluations,” says Mayo. “We released image sets tagged by difficulty along with tools to automatically compute MVT, enabling MVT to be added to existing benchmarks and extended to various applications. These include measuring test set difficulty before deploying real-world systems, discovering neural correlates of image difficulty, and advancing object recognition techniques to close the gap between benchmark and real-world performance.”

    “One of my biggest takeaways is that we now have another dimension to evaluate models on. We want models that are able to recognize any image even if — perhaps especially if — it’s hard for a human to recognize. We’re the first to quantify what this would mean. Our results show that not only is this not the case with today’s state of the art, but also that our current evaluation methods don’t have the ability to tell us when it is the case because standard datasets are so skewed toward easy images,” says Jesse Cummings, an MIT graduate student in electrical engineering and computer science and co-first author with Mayo on the paper.

    From ObjectNet to MVT

    A few years ago, the team behind this project identified a significant challenge in the field of machine learning: Models were struggling with out-of-distribution images, or images that were not well-represented in the training data. Enter ObjectNet, a dataset comprised of images collected from real-life settings. The dataset helped illuminate the performance gap between machine learning models and human recognition abilities, by eliminating spurious correlations present in other benchmarks — for example, between an object and its background. ObjectNet illuminated the gap between the performance of machine vision models on datasets and in real-world applications, encouraging use for many researchers and developers — which subsequently improved model performance.

    Fast forward to the present, and the team has taken their research a step further with MVT. Unlike traditional methods that focus on absolute performance, this new approach assesses how models perform by contrasting their responses to the easiest and hardest images. The study further explored how image difficulty could be explained and tested for similarity to human visual processing. Using metrics like c-score, prediction depth, and adversarial robustness, the team found that harder images are processed differently by networks. “While there are observable trends, such as easier images being more prototypical, a comprehensive semantic explanation of image difficulty continues to elude the scientific community,” says Mayo.

    In the realm of health care, for example, the pertinence of understanding visual complexity becomes even more pronounced. The ability of AI models to interpret medical images, such as X-rays, is subject to the diversity and difficulty distribution of the images. The researchers advocate for a meticulous analysis of difficulty distribution tailored for professionals, ensuring AI systems are evaluated based on expert standards, rather than layperson interpretations.

    Mayo and Cummings are currently looking at neurological underpinnings of visual recognition as well, probing into whether the brain exhibits differential activity when processing easy versus challenging images. The study aims to unravel whether complex images recruit additional brain areas not typically associated with visual processing, hopefully helping demystify how our brains accurately and efficiently decode the visual world.

    Toward human-level performance

    Looking ahead, the researchers are not only focused on exploring ways to enhance AI’s predictive capabilities regarding image difficulty. The team is working on identifying correlations with viewing-time difficulty in order to generate harder or easier versions of images.

    Despite the study’s significant strides, the researchers acknowledge limitations, particularly in terms of the separation of object recognition from visual search tasks. The current methodology does concentrate on recognizing objects, leaving out the complexities introduced by cluttered images.

    “This comprehensive approach addresses the long-standing challenge of objectively assessing progress towards human-level performance in object recognition and opens new avenues for understanding and advancing the field,” says Mayo. “With the potential to adapt the Minimum Viewing Time difficulty metric for a variety of visual tasks, this work paves the way for more robust, human-like performance in object recognition, ensuring that models are truly put to the test and are ready for the complexities of real-world visual understanding.”

    “This is a fascinating study of how human perception can be used to identify weaknesses in the ways AI vision models are typically benchmarked, which overestimate AI performance by concentrating on easy images,” says Alan L. Yuille, Bloomberg Distinguished Professor of Cognitive Science and Computer Science at Johns Hopkins University, who was not involved in the paper. “This will help develop more realistic benchmarks leading not only to improvements to AI but also make fairer comparisons between AI and human perception.” 

    “It’s widely claimed that computer vision systems now outperform humans, and on some benchmark datasets, that’s true,” says Anthropic technical staff member Simon Kornblith PhD ’17, who was also not involved in this work. “However, a lot of the difficulty in those benchmarks comes from the obscurity of what’s in the images; the average person just doesn’t know enough to classify different breeds of dogs. This work instead focuses on images that people can only get right if given enough time. These images are generally much harder for computer vision systems, but the best systems are only a bit worse than humans.”

    Mayo, Cummings, and Xinyu Lin MEng ’22 wrote the paper alongside CSAIL Research Scientist Andrei Barbu, CSAIL Principal Research Scientist Boris Katz, and MIT-IBM Watson AI Lab Principal Researcher Dan Gutfreund. The researchers are affiliates of the MIT Center for Brains, Minds, and Machines.

    The team is presenting their work at the 2023 Conference on Neural Information Processing Systems (NeurIPS). More