Markus Andrews

More stories

  • 150 Shares169 Views

    in Data Management & Statistics

    How machine learning models can amplify inequities in medical diagnosis and treatment

    Prior to receiving a PhD in computer science from MIT in 2017, Marzyeh Ghassemi had already begun to wonder whether the use of AI techniques might enhance the biases that already existed in health care. She was one of the early researchers to take up this issue, and she’s been exploring it ever since. In a new paper, Ghassemi, now an assistant professor in MIT’s Department of Electrical Science and Engineering (EECS), and three collaborators based at the Computer Science and Artificial Intelligence Laboratory, have probed the roots of the disparities that can arise in machine learning, often causing models that perform well overall to falter when it comes to subgroups for which relatively few data have been collected and utilized in the training process. The paper — written by two MIT PhD students, Yuzhe Yang and Haoran Zhang, EECS computer scientist Dina Katabi (the Thuan and Nicole Pham Professor), and Ghassemi — was presented last month at the 40th International Conference on Machine Learning in Honolulu, Hawaii.

    In their analysis, the researchers focused on “subpopulation shifts” — differences in the way machine learning models perform for one subgroup as compared to another. “We want the models to be fair and work equally well for all groups, but instead we consistently observe the presence of shifts among different groups that can lead to inferior medical diagnosis and treatment,” says Yang, who along with Zhang are the two lead authors on the paper. The main point of their inquiry is to determine the kinds of subpopulation shifts that can occur and to uncover the mechanisms behind them so that, ultimately, more equitable models can be developed.

    The new paper “significantly advances our understanding” of the subpopulation shift phenomenon, claims Stanford University computer scientist Sanmi Koyejo. “This research contributes valuable insights for future advancements in machine learning models’ performance on underrepresented subgroups.”

    Camels and cattle

    The MIT group has identified four principal types of shifts — spurious correlations, attribute imbalance, class imbalance, and attribute generalization — which, according to Yang, “have never been put together into a coherent and unified framework. We’ve come up with a single equation that shows you where biases can come from.”

    Biases can, in fact, stem from what the researchers call the class, or from the attribute, or both. To pick a simple example, suppose the task assigned to the machine learning model is to sort images of objects — animals in this case — into two classes: cows and camels. Attributes are descriptors that don’t specifically relate to the class itself. It might turn out, for instance, that all the images used in the analysis show cows standing on grass and camels on sand — grass and sand serving as the attributes here. Given the data available to it, the machine could reach an erroneous conclusion — namely that cows can only be found on grass, not on sand, with the opposite being true for camels. Such a finding would be incorrect, however, giving rise to a spurious correlation, which, Yang explains, is a “special case” among subpopulation shifts — “one in which you have a bias in both the class and the attribute.”

    In a medical setting, one could rely on machine learning models to determine whether a person has pneumonia or not based on an examination of X-ray images. There would be two classes in this situation, one consisting of people who have the lung ailment, another for those who are infection-free. A relatively straightforward case would involve just two attributes: the people getting X-rayed are either female or male. If, in this particular dataset, there were 100 males diagnosed with pneumonia for every one female diagnosed with pneumonia, that could lead to an attribute imbalance, and the model would likely do a better job of correctly detecting pneumonia for a man than for a woman. Similarly, having 1,000 times more healthy (pneumonia-free) subjects than sick ones would lead to a class imbalance, with the model biased toward healthy cases. Attribute generalization is the last shift highlighted in the new study. If your sample contained 100 male patients with pneumonia and zero female subjects with the same illness, you still would like the model to be able to generalize and make predictions about female subjects even though there are no samples in the training data for females with pneumonia.

    The team then took 20 advanced algorithms, designed to carry out classification tasks, and tested them on a dozen datasets to see how they performed across different population groups. They reached some unexpected conclusions: By improving the “classifier,” which is the last layer of the neural network, they were able to reduce the occurrence of spurious correlations and class imbalance, but the other shifts were unaffected. Improvements to the “encoder,” one of the uppermost layers in the neural network, could reduce the problem of attribute imbalance. “However, no matter what we did to the encoder or classifier, we did not see any improvements in terms of attribute generalization,” Yang says, “and we don’t yet know how to address that.”

    Precisely accurate

    There is also the question of assessing how well your model actually works in terms of evenhandedness among different population groups. The metric normally used, called worst-group accuracy or WGA, is based on the assumption that if you can improve the accuracy — of, say, medical diagnosis — for the group that has the worst model performance, you would have improved the model as a whole. “The WGA is considered the gold standard in subpopulation evaluation,” the authors contend, but they made a surprising discovery: boosting worst-group accuracy results in a decrease in what they call “worst-case precision.” In medical decision-making of all sorts, one needs both accuracy — which speaks to the validity of the findings — and precision, which relates to the reliability of the methodology. “Precision and accuracy are both very important metrics in classification tasks, and that is especially true in medical diagnostics,” Yang explains. “You should never trade precision for accuracy. You always need to balance the two.”

    The MIT scientists are putting their theories into practice. In a study they’re conducting with a medical center, they’re looking at public datasets for tens of thousands of patients and hundreds of thousands of chest X-rays, trying to see whether it’s possible for machine learning models to work in an unbiased manner for all populations. That’s still far from the case, even though more awareness has been drawn to this problem, Yang says. “We are finding many disparities across different ages, gender, ethnicity, and intersectional groups.”

    He and his colleagues agree on the eventual goal, which is to achieve fairness in health care among all populations. But before we can reach that point, they maintain, we still need a better understanding of the sources of unfairness and how they permeate our current system. Reforming the system as a whole will not be easy, they acknowledge. In fact, the title of the paper they introduced at the Honolulu conference, “Change is Hard,” gives some indications as to the challenges that they and like-minded researchers face. More

  • 100 Shares119 Views

    in Data Management & Statistics

    Summer research offers a springboard to advanced studies

    Doctoral studies at MIT aren’t a calling for everyone, but they can be for anyone who has had opportunities to discover that science and technology research is their passion and to build the experience and skills to succeed. For Taylor Baum, Josefina Correa Menéndez, and Karla Alejandra Montejo, three graduate students in just one lab of The Picower Institute for Learning and Memory, a pivotal opportunity came via the MIT Summer Research Program in Biology and Neuroscience (MSRP-Bio). When a student finds MSRP-Bio, it helps them find their future in research. 

    In the program, undergraduate STEM majors from outside MIT spend the summer doing full-time research in the departments of Biology, Brain and Cognitive Sciences (BCS), or the Center for Brains, Minds and Machines (CBMM). They gain lab skills, mentoring, preparation for graduate school, and connections that might last a lifetime. Over the last two decades, a total of 215 students from underrepresented minority groups, who are from economically disadvantaged backgrounds, first-generation or nontraditional college students, or students with disabilities have participated in research in BCS or CBMM labs.  

    Like Baum, Correa Menéndez, and Montejo, the vast majority go on to pursue graduate studies, says Diversity and Outreach Coordinator Mandana Sassanfar, who runs the program. For instance, among 91 students who have worked in Picower Institute labs, 81 have completed their undergraduate studies. Of those, 46 enrolled in PhD programs at MIT or other schools such as Cornell, Yale, Stanford, and Princeton universities, and the University of California System. Another 12 have gone to medical school, another seven are in MD/PhD programs, and three have earned master’s degrees. The rest are studying as post-baccalaureates or went straight into the workforce after earning their bachelor’s degree. 

    After participating in the program, Baum, Correa Menéndez, and Montejo each became graduate students in the research group of Emery N. Brown, the Edward Hood Taplin Professor of Computational Neuroscience and Medical Engineering in The Picower Institute and the Institute for Medical Engineering and Science. The lab combines statistical, computational, and experimental neuroscience methods to study how general anesthesia affects the central nervous system to ultimately improve patient care and advance understanding of the brain. Brown says the students have each been doing “off-the-scale” work, in keeping with the excellence he’s seen from MSRP BIO students over the years. For example, on Aug. 10 Baum and Correa Menéndez were honored with MathWorks Fellowships.

    “I think MSRP is fantastic. Mandana does this amazing job of getting students who are quite talented to come to MIT to realize that they can move their game to the next level. They have the capacity to do it. They just need the opportunities,” Brown says. “These students live up to the expectations that you have of them. And now as graduate students, they’re taking on hard problems and they’re solving them.” 

    Paths to PhD studies 

    Pursuing a PhD is hardly a given. Many young students have never considered graduate school or specific fields of study like neuroscience or electrical engineering. But Sassanfar engages students across the country to introduce them to the opportunity MSRP-Bio provides to gain exposure, experience, and mentoring in advanced fields. Every fall, after the program’s students have returned to their undergraduate institutions, she visits schools in places as far flung as Florida, Maryland, Puerto Rico, and Texas and goes to conferences for diverse science communities such as ABRCMS and SACNAS to spread the word. 

    Taylor Baum

    Photo courtesy of Taylor Baum.

    Previous item
    Next item

    When Baum first connected with the program in 2017, she was finding her way at Penn State University. She had been majoring in biology and music composition but had just switched the latter to engineering following a conversation over coffee exposing her to brain-computer interfacing technology, in which detecting brain signals of people with full-body paralysis could improve their quality of life by enabling control of computers or wheelchairs. Baum became enthusiastic about the potential to build similar systems, but as a new engineering student, she struggled to find summer internships and research opportunities. 

    “I got rejected from every single progam except the MIT Center for Brains, Minds and Machines MSRP,” she recalls with a chuckle. 

    Baum thrived in MSRP-Bio, working in Brown’s lab for three successive summers. At each stage, she said, she gained more research skills, experience, and independence. When she graduated, she was sure she wanted to go to graduate school and applied to four of her dream schools. She accepted MIT’s offer to join the Department of Electrical Engineering and Computer Science, where she is co-advised by faculty members there and by Brown. She is now working to develop a system grounded in cardiovascular physiology that can improve blood pressure management. A tool for practicing anesthesiologists, the system automates the dosing of drugs to maintain a patient’s blood pressure at safe levels in the operating room or intensive care unit. 

    More than that, Baum not only is leading an organization advancing STEM education in Puerto Rico, but also is helping to mentor a current MSRP-Bio student in the Brown lab. 

    “MSRP definitely bonds everyone who has participated in it,” Baum says. “If I see anyone who I know participated in MSRP, we could have an immediate conversation. I know that most of us, if we needed help, we’d feel comfortable asking for help from someone from MSRP. With that shared experience, we have a sense of camaraderie, and community.” 

    In fact, a few years ago when a former MSRP-Bio student named Karla Montejo was applying to MIT, Baum provided essential advice and feedback about the application process, Montejo says. Now, as a graduate student, Montejo has become a mentor for the program in her own right, Sassanfar notes. For instance, Montejo serves on program alumni panels that advise new MSRP-Bio students. 

    Karla Alejandra Montejo

    Photo courtesy of Karla Alejandra Montejo.

    Previous item
    Next item

    Montejo’s family immigrated to Miami from Cuba when she was a child. The magnet high school she attended was so new that students were encouraged to help establish the school’s programs. She forged a path into research. 

    “I didn’t even know what research was,” she says. “I wanted to be a doctor, and I thought maybe it would help me on my resume. I thought it would be kind of like shadowing, but no, it was really different. So I got really captured by research when I was in high school.” 

    Despite continuing to pursue research in college at Florida International University, Montejo didn’t get into graduate school on her first attempt because she hadn’t yet learned how to focus her application. But Sassanfar had visited FIU to recruit students and through that relationship Montejo had already gone through MIT’s related Quantitative Methods Workshop (QMW). So Montejo enrolled in MSRP-Bio, working in the CBMM-affiliated lab of Gabriel Kreiman at Boston Children’s Hospital. 

    “I feel like Mandana really helped me out, gave me a break, and the MSRP experience pretty much solidified that I really wanted to come to MIT,” Montejo says. 

    In the QMW, Montejo learned she really liked computational neuroscience, and in Kreiman’s lab she got to try her hand at computational modeling of the cognition involved in making perceptual sense of complex scenes. Montejo realized she wanted to work on more biologically based neuroscience problems. When the summer ended, because she was off the normal graduate school cycle for now, she found a two-year post-baccalaurate program at Mayo Clinic studying the role a brain cell type called astrocytes might have in the Parkinson’s disease treatment deep brain stimulation. 

    When it came time to reapply to graduate schools (with the help of Baum and others in the BCS Application Assistance Program) Montejo applied to MIT and got in, joining the Brown lab. Now she’s working on modeling the role of  metabolic processes in the changing of brain rhythms under anesthesia, taking advantage of how general anesthesia predictably changes brain states. The effects anesthetic drugs have on cell metabolism and the way that ultimately affects levels of consciousness reveals important aspects of how metabolism affects brain circuits and systems. Earlier this month, for instance, Montejo co-led a paper the lab published in The Proceedings of the National Academy of Sciences detailing the neuroscience of a patient’s transition into an especially deep state of unconsciousness called “burst suppression.” 

    Josefina Correa Menendez

    Photo: David Orenstein

    Previous item
    Next item

    A signature of the Brown lab’s work is rigorous statistical analysis and methods, for instance to discern brain arousal states from EEG measures of brain rhythms. A PhD candidate in MIT’s Interdisciplinary Doctoral Program in Statistics, Correa Menéndez is advancing the use of Bayesian hierarchical models for neural data analysis. These statistical models offer a principled way of pooling information across datasets. One of her models can help scientists better understand the way neurons can “spike” with electrical activity when the brain is presented with a stimulus. The other’s power is in discerning critical features such as arousal states of the brain under general anesthesia from electrophysiological recordings. 

    Though she now works with complex equations and computations as a PhD candidate in neuroscience and statistics, Correa Menéndez was mostly interested in music art as a high school student at Academia María Reina in San Juan and then architecture in college at the University of Puerto Rico at Río Piedras. It was discussions at the intersection of epistemology and art during an art theory class that inspired Correa Menéndez to switch her major to biology and to take computer science classes, too. 

    When Sassanfar visited Puerto Rico in 2017, a computer science professor (Patricia Ordóñez) suggested that Correa Menéndez apply for a chance to attend the QMW. She did, and that led her to also participate in MSRP-Bio in the lab of Sherman Fairchild Professor Matt Wilson (a faculty member in BCS, CBMM, and the Picower Institute). She joined in the lab’s studies of how spatial memories are represented in the hippocampus and how the brain makes use of those memories to help understand the world around it. With mentoring from then-postdoc Carmen Varela (now a faculty member at Florida State University), the experience not only exposed her to neuroscience, but also helped her gain skills and experience with lab experiments, building research tools, and conducting statistical analyses. She ended up working in the Wilson lab as a research scholar for a year and began her graduate studies in September 2018.  

    Classes she took with Brown as a research scholar inspired her to join his lab as a graduate student. 

    “Taking the classes with Emery and also doing experiments made me aware of the role of statistics in the scientific process: from the interpretation of results to the analysis and the design of experiments,” she says. “More often than not, in science, statistics becomes this sort of afterthought — this ‘annoying’ thing that people need to do to get their paper published. But statistics as a field is actually a lot more than that. It’s a way of thinking about data. Particularly, Bayesian modeling provides a principled inference framework for combining prior knowledge into a hypothesis that you can test with data.” 

    To be sure, no one starts out with such inspiration about scientific scholarship, but MSRP-Bio helps students find that passion for research and the paths that opens up.   More

  • 63 Shares159 Views

    in Data Management & Statistics

    Embracing the future we need

    When you picture MIT doctoral students taking small PhD courses together, you probably don’t imagine them going on class field trips. But it does happen, sometimes, and one of those trips changed Andy Sun’s career.

    Today, Sun is a faculty member at the MIT Sloan School of Management and a leading global expert on integrating renewable energy into the electric grid. Back in 2007, Sun was an operations research PhD candidate with a diversified academic background: He had studied electrical engineering, quantum computing, and analog computing but was still searching for a doctoral research subject involving energy. 

    One day, as part of a graduate energy class taught by visiting professor Ignacio J. Pérez Arriaga, the students visited the headquarters of ISO-New England, the organization that operates New England’s entire power grid and wholesale electricity market. Suddenly, it hit Sun. His understanding of engineering, used to design and optimize computing systems, could be applied to the grid as a whole, with all its connections, circuitry, and need for efficiency. 

    “The power grids in the U.S. continent are composed of two major interconnections, the Western Interconnection, the Eastern Interconnection, and one minor interconnection, the Texas grid,” Sun says. “Within each interconnection, the power grid is one big machine, essentially. It’s connected by tens of thousands of miles of transmission lines, thousands of generators, and consumers, and if anything is not synchronized, the system may collapse. It’s one of the most complicated engineering systems.”

    And just like that, Sun had a subject he was motivated to pursue. “That’s how I got into this field,” he says. “Taking a field trip.”Sun has barely looked back. He has published dozens of papers about optimizing the flow of intermittent renewable energy through the electricity grid, a major practical issue for grid operators, while also thinking broadly about the future form of the grid and the process of making almost all energy renewable. Sun, who in 2022 rejoined MIT as the Iberdrola-Avangrid Associate Professor in Electric Power Systems, and is also an associate professor of operations research, emphasizes the urgency of rapidly switching to renewables.

    “The decarbonization of our energy system is fundamental,” Sun says. “It will change a lot of things because it has to. We don’t have much time to get there. Two decades, three decades is the window in which we have to get a lot of things done. If you think about how much money will need to be invested, it’s not actually that much. We should embrace this future that we have to get to.”

    Successful operations

    Unexpected as it may have been, Sun’s journey toward being an electricity grid expert was informed by all the stages of his higher education. Sun grew up in China, and received his BA in electronic engineering from Tsinghua University in Beijing, in 2003. He then moved to MIT, joining the Media Lab as a graduate student. Sun intended to study quantum computing but instead began working on analog computer circuit design for Professor Neil Gershenfeld, another person whose worldview influenced Sun.  

    “He had this vision about how optimization is very important in things,” Sun says. “I had never heard of optimization before.” 

    To learn more about it, Sun started taking MIT courses in operations research. “I really enjoyed it, especially the nonlinear optimization course taught by Robert Freund in the Operations Research Center,” he recalls. 

    Sun enjoyed it so much that after a while, he joined MIT’s PhD program in operations research, thanks to the guidance of Freund. Later, he started working with MIT Sloan Professor Dimitri Bertsimas, a leading figure in the field. Still, Sun hadn’t quite nailed down what he wanted to focus on within operations research. Thinking of Sun’s engineering skills, Bertsimas suggested that Sun look for a research topic related to energy. 

    “He wasn’t an expert in energy at that time, but he knew that there are important problems there and encouraged me to go ahead and learn,” Sun says. 

    So it was that Sun found himself in ISO-New England headquarters one day in 2007, finally knowing what he wanted to study, and quickly finding opportunities to start learning from the organization’s experts on electricity markets. By 2011, Sun had finished his MIT PhD dissertation. Based in part on ISO-New England data, the thesis presented new modeling to more efficiently integrate renewable energy into the grid; built some new modeling tools grid operators could use; and developed a way to add fair short-term energy auctions to an efficient grid system.

    The core problem Sun deals with is that, unlike some other sources of electricity, renewables tend to be intermittent, generating power in an uneven pattern over time. That’s not an insurmountable problem for grid operators, but it does require some new approaches. Many of the papers Sun has written focus on precisely how to increasingly draw upon intermittent energy sources while ensuring that the grid’s current level of functionality remains intact. This is also the focus of his 2021 book, co-authored with Antonio J. Conejo, “Robust Optimiziation in Electric Energy Systems.”

    “A major theme of my research is how to achieve the integration of renewables and still operate the system reliably,” Sun says. “You have to keep the balance of supply and demand. This requires many time scales of operation from multidecade planning, to monthly or annual maintenance, to daily operations, down through second-by-second. I work on problems in all these timescales.”

    “I sit in the interface between power engineering and operations research,” Sun says. “I’m not a power engineer, but I sit in this boundary, and I keep the problems in optimization as my motivation.”

    Culture shift

    Sun’s presence on the MIT campus represents a homecoming of sorts. After receiving his doctorate from MIT, Sun spent a year as a postdoc at IBM’s Thomas J. Watson Research Center, then joined the faculty at Georgia Tech, where he remained for a decade. He returned to the Institute in January of 2022.

    “I’m just very excited about the opportunity of being back at MIT,” Sun says. “The MIT Energy Initiative is a such a vibrant place, where many people come together to work on energy. I sit in Sloan, but one very strong point of MIT is there are not many barriers, institutionally. I really look forward to working with colleagues from engineering, Sloan, everywhere, moving forward. We’re moving in the right direction, with a lot of people coming together to break the traditional academic boundaries.” 

    Still, Sun warns that some people may be underestimating the severity of the challenge ahead and the need to implement changes right now. The assets in power grids have long life time, lasting multiple decades. That means investment decisions made now could affect how much clean power is being used a generation from now. 

    “We’re talking about a short timeline, for changing something as huge as how a society fundamentally powers itself with energy,” Sun says. “A lot of that must come from the technology we have today. Renewables are becoming much better and cheaper, so their use has to go up.”

    And that means more people need to work on issues of how to deploy and integrate renewables into everyday life, in the electric grid, transportation, and more. Sun hopes people will increasingly recognize energy as a huge growth area for research and applied work. For instance, when MIT President Sally Kornbluth gave her inaugural address on May 1 this year, she emphasized tackling the climate crisis as her highest priority, something Sun noticed and applauded. 

    “I think the most important thing is the culture,” Sun says. “Bring climate up to the front, and create the platform to encourage people to come together and work on this issue.” More

  • 88 Shares179 Views

    in Data Management & Statistics

    The curse of variety in transportation systems

    Cathy Wu has always delighted in systems that run smoothly. In high school, she designed a project to optimize the best route for getting to class on time. Her research interests and career track are evidence of a propensity for organizing and optimizing, coupled with a strong sense of responsibility to contribute to society instilled by her parents at a young age.

    As an undergraduate at MIT, Wu explored domains like agriculture, energy, and education, eventually homing in on transportation. “Transportation touches each of our lives,” she says. “Every day, we experience the inefficiencies and safety issues as well as the environmental harms associated with our transportation systems. I believe we can and should do better.”

    But doing so is complicated. Consider the long-standing issue of traffic systems control. Wu explains that it is not one problem, but more accurately a family of control problems impacted by variables like time of day, weather, and vehicle type — not to mention the types of sensing and communication technologies used to measure roadway information. Every differentiating factor introduces an exponentially larger set of control problems. There are thousands of control-problem variations and hundreds, if not thousands, of studies and papers dedicated to each problem. Wu refers to the sheer number of variations as the curse of variety — and it is hindering innovation.

    Play video

    “To prove that a new control strategy can be safely deployed on our streets can take years. As time lags, we lose opportunities to improve safety and equity while mitigating environmental impacts. Accelerating this process has huge potential,” says Wu.  

    Which is why she and her group in the MIT Laboratory for Information and Decision Systems are devising machine learning-based methods to solve not just a single control problem or a single optimization problem, but families of control and optimization problems at scale. “In our case, we’re examining emerging transportation problems that people have spent decades trying to solve with classical approaches. It seems to me that we need a different approach.”

    Optimizing intersections

    Currently, Wu’s largest research endeavor is called Project Greenwave. There are many sectors that directly contribute to climate change, but transportation is responsible for the largest share of greenhouse gas emissions — 29 percent, of which 81 percent is due to land transportation. And while much of the conversation around mitigating environmental impacts related to mobility is focused on electric vehicles (EVs), electrification has its drawbacks. EV fleet turnover is time-consuming (“on the order of decades,” says Wu), and limited global access to the technology presents a significant barrier to widespread adoption.

    Wu’s research, on the other hand, addresses traffic control problems by leveraging deep reinforcement learning. Specifically, she is looking at traffic intersections — and for good reason. In the United States alone, there are more than 300,000 signalized intersections where vehicles must stop or slow down before re-accelerating. And every re-acceleration burns fossil fuels and contributes to greenhouse gas emissions.

    Highlighting the magnitude of the issue, Wu says, “We have done preliminary analysis indicating that up to 15 percent of land transportation CO2 is wasted through energy spent idling and re-accelerating at intersections.”

    To date, she and her group have modeled 30,000 different intersections across 10 major metropolitan areas in the United States. That is 30,000 different configurations, roadway topologies (e.g., grade of road or elevation), different weather conditions, and variations in travel demand and fuel mix. Each intersection and its corresponding scenarios represents a unique multi-agent control problem.

    Wu and her team are devising techniques that can solve not just one, but a whole family of problems comprised of tens of thousands of scenarios. Put simply, the idea is to coordinate the timing of vehicles so they arrive at intersections when traffic lights are green, thereby eliminating the start, stop, re-accelerate conundrum. Along the way, they are building an ecosystem of tools, datasets, and methods to enable roadway interventions and impact assessments of strategies to significantly reduce carbon-intense urban driving.

    Play video

    Their collaborator on the project is the Utah Department of Transportation, which Wu says has played an essential role, in part by sharing data and practical knowledge that she and her group otherwise would not have been able to access publicly.

    “I appreciate industry and public sector collaborations,” says Wu. “When it comes to important societal problems, one really needs grounding with practitioners. One needs to be able to hear the perspectives in the field. My interactions with practitioners expand my horizons and help ground my research. You never know when you’ll hear the perspective that is the key to the solution, or perhaps the key to understanding the problem.”

    Finding the best routes

    In a similar vein, she and her research group are tackling large coordination problems. For example, vehicle routing. “Every day, delivery trucks route more than a hundred thousand packages for the city of Boston alone,” says Wu. Accomplishing the task requires, among other things, figuring out which trucks to use, which packages to deliver, and the order in which to deliver them as efficiently as possible. If and when the trucks are electrified, they will need to be charged, adding another wrinkle to the process and further complicating route optimization.

    The vehicle routing problem, and therefore the scope of Wu’s work, extends beyond truck routing for package delivery. Ride-hailing cars may need to pick up objects as well as drop them off; and what if delivery is done by bicycle or drone? In partnership with Amazon, for example, Wu and her team addressed routing and path planning for hundreds of robots (up to 800) in their warehouses.

    Every variation requires custom heuristics that are expensive and time-consuming to develop. Again, this is really a family of problems — each one complicated, time-consuming, and currently unsolved by classical techniques — and they are all variations of a central routing problem. The curse of variety meets operations and logistics.

    By combining classical approaches with modern deep-learning methods, Wu is looking for a way to automatically identify heuristics that can effectively solve all of these vehicle routing problems. So far, her approach has proved successful.

    “We’ve contributed hybrid learning approaches that take existing solution methods for small problems and incorporate them into our learning framework to scale and accelerate that existing solver for large problems. And we’re able to do this in a way that can automatically identify heuristics for specialized variations of the vehicle routing problem.” The next step, says Wu, is applying a similar approach to multi-agent robotics problems in automated warehouses.

    Wu and her group are making big strides, in part due to their dedication to use-inspired basic research. Rather than applying known methods or science to a problem, they develop new methods, new science, to address problems. The methods she and her team employ are necessitated by societal problems with practical implications. The inspiration for the approach? None other than Louis Pasteur, who described his research style in a now-famous article titled “Pasteur’s Quadrant.” Anthrax was decimating the sheep population, and Pasteur wanted to better understand why and what could be done about it. The tools of the time could not solve the problem, so he invented a new field, microbiology, not out of curiosity but out of necessity. More

  • 63 Shares169 Views

    in Data Management & Statistics

    Making sense of cell fate

    Despite the proliferation of novel therapies such as immunotherapy or targeted therapies, radiation and chemotherapy remain the frontline treatment for cancer patients. About half of all patients still receive radiation and 60-80 percent receive chemotherapy.

    Both radiation and chemotherapy work by damaging DNA, taking advantage of a vulnerability specific to cancer cells. Healthy cells are more likely to survive radiation and chemotherapy since their mechanisms for identifying and repairing DNA damage are intact. In cancer cells, these repair mechanisms are compromised by mutations. When cancer cells cannot adequately respond to the DNA damage caused by radiation and chemotherapy, ideally, they undergo apoptosis or die by other means.

    However, there is another fate for cells after DNA damage: senescence — a state where cells survive, but stop dividing. Senescent cells’ DNA has not been damaged enough to induce apoptosis but is too damaged to support cell division. While senescent cancer cells themselves are unable to proliferate and spread, they are bad actors in the fight against cancer because they seem to enable other cancer cells to develop more aggressively. Although a cancer cell’s fate is not apparent until a few days after treatment, the decision to survive, die, or enter senescence is made much earlier. But, precisely when and how that decision is made has not been well understood.

    In an open-access study of ovarian and osteosarcoma cancer cells appearing July 19 in Cell Systems, MIT researchers show that cell signaling proteins commonly associated with cell proliferation and apoptosis instead commit cancer cells to senescence within 12 hours of treatment with low doses of certain kinds of chemotherapy.

    “When it comes to treating cancer, this study underscores that it’s important not to think too linearly about cell signaling,” says Michael Yaffe, who is a David H. Koch Professor of Science at MIT, the director of the MIT Center for Precision Cancer Medicine, a member of MIT’s Koch Institute for Integrative Cancer Research, and the senior author of the study. “If you assume that a particular treatment will always affect cancer cell signaling in the same way — you may be setting yourself up for many surprises, and treating cancers with the wrong combination of drugs.”

    Using a combination of experiments with cancer cells and computational modeling, the team investigated the cell signaling mechanisms that prompt cancer cells to enter senescence after treatment with a commonly used anti-cancer agent. Their efforts singled out two protein kinases and a component of the AP-1 transcription factor complex as highly associated with the induction of senescence after DNA damage, despite the well-established roles for all of these molecules in promoting cell proliferation in cancer.

    The researchers treated cancer cells with low and high doses of doxorubicin, a chemotherapy that interferes with the function with topoisomerase II, an enzyme that breaks and then repairs DNA strands during replication to fix tangles and other topological problems.

    By measuring the effects of DNA damage on single cells at several time points ranging from six hours to four days after the initial exposure, the team created two datasets. In one dataset, the researchers tracked cell fate over time. For the second set, researchers measured relative cell signaling activity levels across a variety of proteins associated with responses to DNA damage or cellular stress, determination of cell fate, and progress through cell growth and division.

    The two datasets were used to build a computational model that identifies correlations between time, dosage, signal, and cell fate. The model identified the activities of the MAP kinases Erk and JNK, and the transcription factor c-Jun as key components of the AP-1 protein likewise understood to involved in the induction of senescence. The researchers then validated these computational findings by showing that inhibition of JNK and Erk after DNA damage successfully prevented cells from entering senescence.

    The researchers leveraged JNK and Erk inhibition to pinpoint exactly when cells made the decision to enter senescence. Surprisingly, they found that the decision to enter senescence was made within 12 hours of DNA damage, even though it took days to actually see the senescent cells accumulate. The team also found that with the passage of more time, these MAP kinases took on a different function: promoting the secretion of proinflammatory proteins called cytokines that are responsible for making other cancer cells proliferate and develop resistance to chemotherapy.

    “Proteins like cytokines encourage ‘bad behavior’ in neighboring tumor cells that lead to more aggressive cancer progression,” says Tatiana Netterfield, a graduate student in the Yaffe lab and the lead author of the study. “Because of this, it is thought that senescent cells that stay near the tumor for long periods of time are detrimental to treating cancer.”

    This study’s findings apply to cancer cells treated with a commonly used type of chemotherapy that stalls DNA replication after repair. But more broadly, the study emphasizes that “when treating cancer, it’s extremely important to understand the molecular characteristics of cancer cells and the contextual factors such as time and dosing that determine cell fate,” explains Netterfield.

    The study, however, has more immediate implications for treatments that are already in use. One class of Erk inhibitors, MEK inhibitors, are used in the clinic with the expectation that they will curb cancer growth.

    “We must be cautious about administering MEK inhibitors together with chemotherapies,” says Yaffe. “The combination may have the unintended effect of driving cells into proliferation, rather than senescence.”

    In future work, the team will perform studies to understand how and why individual cells choose to proliferate instead of enter senescence. Additionally, the team is employing next-generation sequencing to understand which genes c-Jun is regulating in order to push cells toward senescence.

    This study was funded, in part, by the Charles and Marjorie Holloway Foundation and the MIT Center for Precision Cancer Medicine. More

  • 113 Shares159 Views

    in Data Management & Statistics

    A simpler method for learning to control a robot

    Researchers from MIT and Stanford University have devised a new machine-learning approach that could be used to control a robot, such as a drone or autonomous vehicle, more effectively and efficiently in dynamic environments where conditions can change rapidly.

    This technique could help an autonomous vehicle learn to compensate for slippery road conditions to avoid going into a skid, allow a robotic free-flyer to tow different objects in space, or enable a drone to closely follow a downhill skier despite being buffeted by strong winds.

    The researchers’ approach incorporates certain structure from control theory into the process for learning a model in such a way that leads to an effective method of controlling complex dynamics, such as those caused by impacts of wind on the trajectory of a flying vehicle. One way to think about this structure is as a hint that can help guide how to control a system.

    “The focus of our work is to learn intrinsic structure in the dynamics of the system that can be leveraged to design more effective, stabilizing controllers,” says Navid Azizan, the Esther and Harold E. Edgerton Assistant Professor in the MIT Department of Mechanical Engineering and the Institute for Data, Systems, and Society (IDSS), and a member of the Laboratory for Information and Decision Systems (LIDS). “By jointly learning the system’s dynamics and these unique control-oriented structures from data, we’re able to naturally create controllers that function much more effectively in the real world.”

    Using this structure in a learned model, the researchers’ technique immediately extracts an effective controller from the model, as opposed to other machine-learning methods that require a controller to be derived or learned separately with additional steps. With this structure, their approach is also able to learn an effective controller using fewer data than other approaches. This could help their learning-based control system achieve better performance faster in rapidly changing environments.

    “This work tries to strike a balance between identifying structure in your system and just learning a model from data,” says lead author Spencer M. Richards, a graduate student at Stanford University. “Our approach is inspired by how roboticists use physics to derive simpler models for robots. Physical analysis of these models often yields a useful structure for the purposes of control — one that you might miss if you just tried to naively fit a model to data. Instead, we try to identify similarly useful structure from data that indicates how to implement your control logic.”

    Additional authors of the paper are Jean-Jacques Slotine, professor of mechanical engineering and of brain and cognitive sciences at MIT, and Marco Pavone, associate professor of aeronautics and astronautics at Stanford. The research will be presented at the International Conference on Machine Learning (ICML).

    Learning a controller

    Determining the best way to control a robot to accomplish a given task can be a difficult problem, even when researchers know how to model everything about the system.

    A controller is the logic that enables a drone to follow a desired trajectory, for example. This controller would tell the drone how to adjust its rotor forces to compensate for the effect of winds that can knock it off a stable path to reach its goal.

    This drone is a dynamical system — a physical system that evolves over time. In this case, its position and velocity change as it flies through the environment. If such a system is simple enough, engineers can derive a controller by hand. 

    Modeling a system by hand intrinsically captures a certain structure based on the physics of the system. For instance, if a robot were modeled manually using differential equations, these would capture the relationship between velocity, acceleration, and force. Acceleration is the rate of change in velocity over time, which is determined by the mass of and forces applied to the robot.

    But often the system is too complex to be exactly modeled by hand. Aerodynamic effects, like the way swirling wind pushes a flying vehicle, are notoriously difficult to derive manually, Richards explains. Researchers would instead take measurements of the drone’s position, velocity, and rotor speeds over time, and use machine learning to fit a model of this dynamical system to the data. But these approaches typically don’t learn a control-based structure. This structure is useful in determining how to best set the rotor speeds to direct the motion of the drone over time.

    Once they have modeled the dynamical system, many existing approaches also use data to learn a separate controller for the system.

    “Other approaches that try to learn dynamics and a controller from data as separate entities are a bit detached philosophically from the way we normally do it for simpler systems. Our approach is more reminiscent of deriving models by hand from physics and linking that to control,” Richards says.

    Identifying structure

    The team from MIT and Stanford developed a technique that uses machine learning to learn the dynamics model, but in such a way that the model has some prescribed structure that is useful for controlling the system.

    With this structure, they can extract a controller directly from the dynamics model, rather than using data to learn an entirely separate model for the controller.

    “We found that beyond learning the dynamics, it’s also essential to learn the control-oriented structure that supports effective controller design. Our approach of learning state-dependent coefficient factorizations of the dynamics has outperformed the baselines in terms of data efficiency and tracking capability, proving to be successful in efficiently and effectively controlling the system’s trajectory,” Azizan says. 

    When they tested this approach, their controller closely followed desired trajectories, outpacing all the baseline methods. The controller extracted from their learned model nearly matched the performance of a ground-truth controller, which is built using the exact dynamics of the system.

    “By making simpler assumptions, we got something that actually worked better than other complicated baseline approaches,” Richards adds.

    The researchers also found that their method was data-efficient, which means it achieved high performance even with few data. For instance, it could effectively model a highly dynamic rotor-driven vehicle using only 100 data points. Methods that used multiple learned components saw their performance drop much faster with smaller datasets.

    This efficiency could make their technique especially useful in situations where a drone or robot needs to learn quickly in rapidly changing conditions.

    Plus, their approach is general and could be applied to many types of dynamical systems, from robotic arms to free-flying spacecraft operating in low-gravity environments.

    In the future, the researchers are interested in developing models that are more physically interpretable, and that would be able to identify very specific information about a dynamical system, Richards says. This could lead to better-performing controllers.

    “Despite its ubiquity and importance, nonlinear feedback control remains an art, making it especially suitable for data-driven and learning-based methods. This paper makes a significant contribution to this area by proposing a method that jointly learns system dynamics, a controller, and control-oriented structure,” says Nikolai Matni, an assistant professor in the Department of Electrical and Systems Engineering at the University of Pennsylvania, who was not involved with this work. “What I found particularly exciting and compelling was the integration of these components into a joint learning algorithm, such that control-oriented structure acts as an inductive bias in the learning process. The result is a data-efficient learning process that outputs dynamic models that enjoy intrinsic structure that enables effective, stable, and robust control. While the technical contributions of the paper are excellent themselves, it is this conceptual contribution that I view as most exciting and significant.”

    This research is supported, in part, by the NASA University Leadership Initiative and the Natural Sciences and Engineering Research Council of Canada. More

  • 138 Shares129 Views

    in Data Management & Statistics

    A new dataset of Arctic images will spur artificial intelligence research

    As the U.S. Coast Guard (USCG) icebreaker Healy takes part in a voyage across the North Pole this summer, it is capturing images of the Arctic to further the study of this rapidly changing region. Lincoln Laboratory researchers installed a camera system aboard the Healy while at port in Seattle before it embarked on a three-month science mission on July 11. The resulting dataset, which will be one of the first of its kind, will be used to develop artificial intelligence tools that can analyze Arctic imagery.

    “This dataset not only can help mariners navigate more safely and operate more efficiently, but also help protect our nation by providing critical maritime domain awareness and an improved understanding of how AI analysis can be brought to bear in this challenging and unique environment,” says Jo Kurucar, a researcher in Lincoln Laboratory’s AI Software Architectures and Algorithms Group, which led this project.

    As the planet warms and sea ice melts, Arctic passages are opening up to more traffic, both to military vessels and ships conducting illegal fishing. These movements may pose national security challenges to the United States. The opening Arctic also leaves questions about how its climate, wildlife, and geography are changing.

    Today, very few imagery datasets of the Arctic exist to study these changes. Overhead images from satellites or aircraft can only provide limited information about the environment. An outward-looking camera attached to a ship can capture more details of the setting and different angles of objects, such as other ships, in the scene. These types of images can then be used to train AI computer-vision tools, which can help the USCG plan naval missions and automate analysis. According to Kurucar, USCG assets in the Arctic are spread thin and can benefit greatly from AI tools, which can act as a force multiplier.

    The Healy is the USCG’s largest and most technologically advanced icebreaker. Given its current mission, it was a fitting candidate to be equipped with a new sensor to gather this dataset. The laboratory research team collaborated with the USCG Research and Development Center to determine the sensor requirements. Together, they developed the Cold Region Imaging and Surveillance Platform (CRISP).

    “Lincoln Laboratory has an excellent relationship with the Coast Guard, especially with the Research and Development Center. Over a decade, we’ve established ties that enabled the deployment of the CRISP system,” says Amna Greaves, the CRISP project lead and an assistant leader in the AI Software Architectures and Algorithms Group. “We have strong ties not only because of the USCG veterans working at the laboratory and in our group, but also because our technology missions are complementary. Today it was deploying infrared sensing in the Arctic; tomorrow it could be operating quadruped robot dogs on a fast-response cutter.”

    The CRISP system comprises a long-wave infrared camera, manufactured by Teledyne FLIR (for forward-looking infrared), that is designed for harsh maritime environments. The camera can stabilize itself during rough seas and image in complete darkness, fog, and glare. It is paired with a GPS-enabled time-synchronized clock and a network video recorder to record both video and still imagery along with GPS-positional data.  

    The camera is mounted at the front of the ship’s fly bridge, and the electronics are housed in a ruggedized rack on the bridge. The system can be operated manually from the bridge or be placed into an autonomous surveillance mode, in which it slowly pans back and forth, recording 15 minutes of video every three hours and a still image once every 15 seconds.

    “The installation of the equipment was a unique and fun experience. As with any good project, our expectations going into the install did not meet reality,” says Michael Emily, the project’s IT systems administrator who traveled to Seattle for the install. Working with the ship’s crew, the laboratory team had to quickly adjust their route for running cables from the camera to the observation station after they discovered that the expected access points weren’t in fact accessible. “We had 100-foot cables made for this project just in case of this type of scenario, which was a good thing because we only had a few inches to spare,” Emily says.

    The CRISP project team plans to publicly release the dataset, anticipated to be about 4 terabytes in size, once the USCG science mission concludes in the fall.

    The goal in releasing the dataset is to enable the wider research community to develop better tools for those operating in the Arctic, especially as this region becomes more navigable. “Collecting and publishing the data allows for faster and greater progress than what we could accomplish on our own,” Kurucar adds. “It also enables the laboratory to engage in more advanced AI applications while others make more incremental advances using the dataset.”

    On top of providing the dataset, the laboratory team plans to provide a baseline object-detection model, from which others can make progress on their own models. More advanced AI applications planned for development are classifiers for specific objects in the scene and the ability to identify and track objects across images.

    Beyond assisting with USCG missions, this project could create an influential dataset for researchers looking to apply AI to data from the Arctic to help combat climate change, says Paul Metzger, who leads the AI Software Architectures and Algorithms Group.

    Metzger adds that the group was honored to be a part of this project and is excited to see the advances that come from applying AI to novel challenges facing the United States: “I’m extremely proud of how our group applies AI to the highest-priority challenges in our nation, from predicting outbreaks of Covid-19 and assisting the U.S. European Command in their support of Ukraine to now employing AI in the Arctic for maritime awareness.”

    Once the dataset is available, it will be free to download on the Lincoln Laboratory dataset website. More

  • 88 Shares109 Views

    in Data Management & Statistics

    A faster way to teach a robot

    Imagine purchasing a robot to perform household tasks. This robot was built and trained in a factory on a certain set of tasks and has never seen the items in your home. When you ask it to pick up a mug from your kitchen table, it might not recognize your mug (perhaps because this mug is painted with an unusual image, say, of MIT’s mascot, Tim the Beaver). So, the robot fails.

    “Right now, the way we train these robots, when they fail, we don’t really know why. So you would just throw up your hands and say, ‘OK, I guess we have to start over.’ A critical component that is missing from this system is enabling the robot to demonstrate why it is failing so the user can give it feedback,” says Andi Peng, an electrical engineering and computer science (EECS) graduate student at MIT.

    Peng and her collaborators at MIT, New York University, and the University of California at Berkeley created a framework that enables humans to quickly teach a robot what they want it to do, with a minimal amount of effort.

    When a robot fails, the system uses an algorithm to generate counterfactual explanations that describe what needed to change for the robot to succeed. For instance, maybe the robot would have been able to pick up the mug if the mug were a certain color. It shows these counterfactuals to the human and asks for feedback on why the robot failed. Then the system utilizes this feedback and the counterfactual explanations to generate new data it uses to fine-tune the robot.

    Fine-tuning involves tweaking a machine-learning model that has already been trained to perform one task, so it can perform a second, similar task.

    The researchers tested this technique in simulations and found that it could teach a robot more efficiently than other methods. The robots trained with this framework performed better, while the training process consumed less of a human’s time.

    This framework could help robots learn faster in new environments without requiring a user to have technical knowledge. In the long run, this could be a step toward enabling general-purpose robots to efficiently perform daily tasks for the elderly or individuals with disabilities in a variety of settings.

    Peng, the lead author, is joined by co-authors Aviv Netanyahu, an EECS graduate student; Mark Ho, an assistant professor at the Stevens Institute of Technology; Tianmin Shu, an MIT postdoc; Andreea Bobu, a graduate student at UC Berkeley; and senior authors Julie Shah, an MIT professor of aeronautics and astronautics and the director of the Interactive Robotics Group in the Computer Science and Artificial Intelligence Laboratory (CSAIL), and Pulkit Agrawal, a professor in CSAIL. The research will be presented at the International Conference on Machine Learning.

    On-the-job training

    Robots often fail due to distribution shift — the robot is presented with objects and spaces it did not see during training, and it doesn’t understand what to do in this new environment.

    One way to retrain a robot for a specific task is imitation learning. The user could demonstrate the correct task to teach the robot what to do. If a user tries to teach a robot to pick up a mug, but demonstrates with a white mug, the robot could learn that all mugs are white. It may then fail to pick up a red, blue, or “Tim-the-Beaver-brown” mug.

    Training a robot to recognize that a mug is a mug, regardless of its color, could take thousands of demonstrations.

    “I don’t want to have to demonstrate with 30,000 mugs. I want to demonstrate with just one mug. But then I need to teach the robot so it recognizes that it can pick up a mug of any color,” Peng says.

    To accomplish this, the researchers’ system determines what specific object the user cares about (a mug) and what elements aren’t important for the task (perhaps the color of the mug doesn’t matter). It uses this information to generate new, synthetic data by changing these “unimportant” visual concepts. This process is known as data augmentation.

    The framework has three steps. First, it shows the task that caused the robot to fail. Then it collects a demonstration from the user of the desired actions and generates counterfactuals by searching over all features in the space that show what needed to change for the robot to succeed.

    The system shows these counterfactuals to the user and asks for feedback to determine which visual concepts do not impact the desired action. Then it uses this human feedback to generate many new augmented demonstrations.

    In this way, the user could demonstrate picking up one mug, but the system would produce demonstrations showing the desired action with thousands of different mugs by altering the color. It uses these data to fine-tune the robot.

    Creating counterfactual explanations and soliciting feedback from the user are critical for the technique to succeed, Peng says.

    From human reasoning to robot reasoning

    Because their work seeks to put the human in the training loop, the researchers tested their technique with human users. They first conducted a study in which they asked people if counterfactual explanations helped them identify elements that could be changed without affecting the task.

    “It was so clear right off the bat. Humans are so good at this type of counterfactual reasoning. And this counterfactual step is what allows human reasoning to be translated into robot reasoning in a way that makes sense,” she says.

    Then they applied their framework to three simulations where robots were tasked with: navigating to a goal object, picking up a key and unlocking a door, and picking up a desired object then placing it on a tabletop. In each instance, their method enabled the robot to learn faster than with other techniques, while requiring fewer demonstrations from users.

    Moving forward, the researchers hope to test this framework on real robots. They also want to focus on reducing the time it takes the system to create new data using generative machine-learning models.

    “We want robots to do what humans do, and we want them to do it in a semantically meaningful way. Humans tend to operate in this abstract space, where they don’t think about every single property in an image. At the end of the day, this is really about enabling a robot to learn a good, human-like representation at an abstract level,” Peng says.

    This research is supported, in part, by a National Science Foundation Graduate Research Fellowship, Open Philanthropy, an Apple AI/ML Fellowship, Hyundai Motor Corporation, the MIT-IBM Watson AI Lab, and the National Science Foundation Institute for Artificial Intelligence and Fundamental Interactions. More