MIT Schwarzman College of Computing

Subterms

    More stories

    • 138 Shares129 Views

      in Data Management & Statistics

      Study: AI models fail to reproduce human judgements about rule violations

      by

      In an effort to improve fairness or reduce backlogs, machine-learning models are sometimes designed to mimic human decision making, such as deciding whether social media posts violate toxic content policies.

      But researchers from MIT and elsewhere have found that these models often do not replicate human decisions about rule violations. If models are not trained with the right data, they are likely to make different, often harsher judgements than humans would.

      In this case, the “right” data are those that have been labeled by humans who were explicitly asked whether items defy a certain rule. Training involves showing a machine-learning model millions of examples of this “normative data” so it can learn a task.

      But data used to train machine-learning models are typically labeled descriptively — meaning humans are asked to identify factual features, such as, say, the presence of fried food in a photo. If “descriptive data” are used to train models that judge rule violations, such as whether a meal violates a school policy that prohibits fried food, the models tend to over-predict rule violations.

      This drop in accuracy could have serious implications in the real world. For instance, if a descriptive model is used to make decisions about whether an individual is likely to reoffend, the researchers’ findings suggest it may cast stricter judgements than a human would, which could lead to higher bail amounts or longer criminal sentences.

      “I think most artificial intelligence/machine-learning researchers assume that the human judgements in data and labels are biased, but this result is saying something worse. These models are not even reproducing already-biased human judgments because the data they’re being trained on has a flaw: Humans would label the features of images and text differently if they knew those features would be used for a judgment. This has huge ramifications for machine learning systems in human processes,” says Marzyeh Ghassemi, an assistant professor and head of the Healthy ML Group in the Computer Science and Artificial Intelligence Laboratory (CSAIL).

      Ghassemi is senior author of a new paper detailing these findings, which was published today in Science Advances. Joining her on the paper are lead author Aparna Balagopalan, an electrical engineering and computer science graduate student; David Madras, a graduate student at the University of Toronto; David H. Yang, a former graduate student who is now co-founder of ML Estimation; Dylan Hadfield-Menell, an MIT assistant professor; and Gillian K. Hadfield, Schwartz Reisman Chair in Technology and Society and professor of law at the University of Toronto.

      Labeling discrepancy

      This study grew out of a different project that explored how a machine-learning model can justify its predictions. As they gathered data for that study, the researchers noticed that humans sometimes give different answers if they are asked to provide descriptive or normative labels about the same data.

      To gather descriptive labels, researchers ask labelers to identify factual features — does this text contain obscene language? To gather normative labels, researchers give labelers a rule and ask if the data violates that rule — does this text violate the platform’s explicit language policy?

      Surprised by this finding, the researchers launched a user study to dig deeper. They gathered four datasets to mimic different policies, such as a dataset of dog images that could be in violation of an apartment’s rule against aggressive breeds. Then they asked groups of participants to provide descriptive or normative labels.

      In each case, the descriptive labelers were asked to indicate whether three factual features were present in the image or text, such as whether the dog appears aggressive. Their responses were then used to craft judgements. (If a user said a photo contained an aggressive dog, then the policy was violated.) The labelers did not know the pet policy. On the other hand, normative labelers were given the policy prohibiting aggressive dogs, and then asked whether it had been violated by each image, and why.

      The researchers found that humans were significantly more likely to label an object as a violation in the descriptive setting. The disparity, which they computed using the absolute difference in labels on average, ranged from 8 percent on a dataset of images used to judge dress code violations to 20 percent for the dog images.

      “While we didn’t explicitly test why this happens, one hypothesis is that maybe how people think about rule violations is different from how they think about descriptive data. Generally, normative decisions are more lenient,” Balagopalan says.

      Yet data are usually gathered with descriptive labels to train a model for a particular machine-learning task. These data are often repurposed later to train different models that perform normative judgements, like rule violations.

      Training troubles

      To study the potential impacts of repurposing descriptive data, the researchers trained two models to judge rule violations using one of their four data settings. They trained one model using descriptive data and the other using normative data, and then compared their performance.

      They found that if descriptive data are used to train a model, it will underperform a model trained to perform the same judgements using normative data. Specifically, the descriptive model is more likely to misclassify inputs by falsely predicting a rule violation. And the descriptive model’s accuracy was even lower when classifying objects that human labelers disagreed about.

      “This shows that the data do really matter. It is important to match the training context to the deployment context if you are training models to detect if a rule has been violated,” Balagopalan says.

      It can be very difficult for users to determine how data have been gathered; this information can be buried in the appendix of a research paper or not revealed by a private company, Ghassemi says.

      Improving dataset transparency is one way this problem could be mitigated. If researchers know how data were gathered, then they know how those data should be used. Another possible strategy is to fine-tune a descriptively trained model on a small amount of normative data. This idea, known as transfer learning, is something the researchers want to explore in future work.

      They also want to conduct a similar study with expert labelers, like doctors or lawyers, to see if it leads to the same label disparity.

      “The way to fix this is to transparently acknowledge that if we want to reproduce human judgment, we must only use data that were collected in that setting. Otherwise, we are going to end up with systems that are going to have extremely harsh moderations, much harsher than what humans would do. Humans would see nuance or make another distinction, whereas these models don’t,” Ghassemi says.

      This research was funded, in part, by the Schwartz Reisman Institute for Technology and Society, Microsoft Research, the Vector Institute, and a Canada Research Council Chain. More

    • 113 Shares119 Views

      in Data Management & Statistics

      Researchers create a tool for accurately simulating complex systems

      by

      Researchers often use simulations when designing new algorithms, since testing ideas in the real world can be both costly and risky. But since it’s impossible to capture every detail of a complex system in a simulation, they typically collect a small amount of real data that they replay while simulating the components they want to study.

      Known as trace-driven simulation (the small pieces of real data are called traces), this method sometimes results in biased outcomes. This means researchers might unknowingly choose an algorithm that is not the best one they evaluated, and which will perform worse on real data than the simulation predicted that it should.

      MIT researchers have developed a new method that eliminates this source of bias in trace-driven simulation. By enabling unbiased trace-driven simulations, the new technique could help researchers design better algorithms for a variety of applications, including improving video quality on the internet and increasing the performance of data processing systems.

      The researchers’ machine-learning algorithm draws on the principles of causality to learn how the data traces were affected by the behavior of the system. In this way, they can replay the correct, unbiased version of the trace during the simulation.

      When compared to a previously developed trace-driven simulator, the researchers’ simulation method correctly predicted which newly designed algorithm would be best for video streaming — meaning the one that led to less rebuffering and higher visual quality. Existing simulators that do not account for bias would have pointed researchers to a worse-performing algorithm.

      “Data are not the only thing that matter. The story behind how the data are generated and collected is also important. If you want to answer a counterfactual question, you need to know the underlying data generation story so you only intervene on those things that you really want to simulate,” says Arash Nasr-Esfahany, an electrical engineering and computer science (EECS) graduate student and co-lead author of a paper on this new technique.

      He is joined on the paper by co-lead authors and fellow EECS graduate students Abdullah Alomar and Pouya Hamadanian; recent graduate student Anish Agarwal PhD ’21; and senior authors Mohammad Alizadeh, an associate professor of electrical engineering and computer science; and Devavrat Shah, the Andrew and Erna Viterbi Professor in EECS and a member of the Institute for Data, Systems, and Society and of the Laboratory for Information and Decision Systems. The research was recently presented at the USENIX Symposium on Networked Systems Design and Implementation.

      Specious simulations

      The MIT researchers studied trace-driven simulation in the context of video streaming applications.

      In video streaming, an adaptive bitrate algorithm continually decides the video quality, or bitrate, to transfer to a device based on real-time data on the user’s bandwidth. To test how different adaptive bitrate algorithms impact network performance, researchers can collect real data from users during a video stream for a trace-driven simulation.

      They use these traces to simulate what would have happened to network performance had the platform used a different adaptive bitrate algorithm in the same underlying conditions.

      Researchers have traditionally assumed that trace data are exogenous, meaning they aren’t affected by factors that are changed during the simulation. They would assume that, during the period when they collected the network performance data, the choices the bitrate adaptation algorithm made did not affect those data.

      But this is often a false assumption that results in biases about the behavior of new algorithms, making the simulation invalid, Alizadeh explains.

      “We recognized, and others have recognized, that this way of doing simulation can induce errors. But I don’t think people necessarily knew how significant those errors could be,” he says.

      To develop a solution, Alizadeh and his collaborators framed the issue as a causal inference problem. To collect an unbiased trace, one must understand the different causes that affect the observed data. Some causes are intrinsic to a system, while others are affected by the actions being taken.

      In the video streaming example, network performance is affected by the choices the bitrate adaptation algorithm made — but it’s also affected by intrinsic elements, like network capacity.

      “Our task is to disentangle these two effects, to try to understand what aspects of the behavior we are seeing are intrinsic to the system and how much of what we are observing is based on the actions that were taken. If we can disentangle these two effects, then we can do unbiased simulations,” he says.

      Learning from data

      But researchers often cannot directly observe intrinsic properties. This is where the new tool, called CausalSim, comes in. The algorithm can learn the underlying characteristics of a system using only the trace data.

      CausalSim takes trace data that were collected through a randomized control trial, and estimates the underlying functions that produced those data. The model tells the researchers, under the exact same underlying conditions that a user experienced, how a new algorithm would change the outcome.

      Using a typical trace-driven simulator, bias might lead a researcher to select a worse-performing algorithm, even though the simulation indicates it should be better. CausalSim helps researchers select the best algorithm that was tested.

      The MIT researchers observed this in practice. When they used CausalSim to design an improved bitrate adaptation algorithm, it led them to select a new variant that had a stall rate that was nearly 1.4 times lower than a well-accepted competing algorithm, while achieving the same video quality. The stall rate is the amount of time a user spent rebuffering the video.

      By contrast, an expert-designed trace-driven simulator predicted the opposite. It indicated that this new variant should cause a stall rate that was nearly 1.3 times higher. The researchers tested the algorithm on real-world video streaming and confirmed that CausalSim was correct.

      “The gains we were getting in the new variant were very close to CausalSim’s prediction, while the expert simulator was way off. This is really exciting because this expert-designed simulator has been used in research for the past decade. If CausalSim can so clearly be better than this, who knows what we can do with it?” says Hamadanian.

      During a 10-month experiment, CausalSim consistently improved simulation accuracy, resulting in algorithms that made about half as many errors as those designed using baseline methods.

      In the future, the researchers want to apply CausalSim to situations where randomized control trial data are not available or where it is especially difficult to recover the causal dynamics of the system. They also want to explore how to design and monitor systems to make them more amenable to causal analysis. More

    • 88 Shares99 Views

      in Data Management & Statistics

      Study: Covid-19 has reduced diverse urban interactions

      by

      The Covid-19 pandemic has reduced how often urban residents intersect with people from different income brackets, according to a new study led by MIT researchers.

      Examining the movement of people in four U.S. cities before and after the onset of the pandemic, the study found a 15 to 30 percent decrease in the number of visits residents were making to areas that are socioeconomically different than their own. In turn, this has reduced people’s opportunities to interact with others from varied social and economic spheres.

      “Income diversity of urban encounters decreased during the pandemic, and not just in the lockdown stages,” says Takahiro Yabe, a postdoc at the Media Lab and co-author of a newly published paper detailing the study’s results. “It decreased in the long term as well, after mobility patterns recovered.”

      Indeed, the study found a large immediate dropoff in urban movement in the spring of 2020, when new policies temporarily shuttered many types of institutions and businesses in the U.S. and much of the world due to the emergence of the deadly Covid-19 virus. But even after such restrictions were lifted and the overall amount of urban movement approached prepandemic levels, movement patterns within cities have narrowed; people now visit fewer places.

      “We see that changes like working from home, less exploration, more online shopping, all these behaviors add up,” says Esteban Moro, a research scientist at MIT’s Sociotechnical Systems Research Center (SSRC) and another of the paper’s co-authors. “Working from home is amazing and shopping online is great, but we are not seeing each other at the rates we were before.”

      The paper, “Behavioral changes during the Covid-19 pandemic decreased income diversity of urban encounters,” appears in Nature Communications. The co-authors are Yabe; Bernardo García Bulle Bueno, a doctoral candidate at MIT’s Institute for Data, Systems, and Society (IDSS); Xiaowen Dong, an associate professor at Oxford University; Alex Pentland, professor of media arts and sciences at MIT and the Toshiba Professor at the Media Lab; and Moro, who is also an associate professor at the University Carlos III of Madrid.

      A decline in exploration

      To conduct the study, the researchers examined anonymized cellphone data from 1 million users over a three-year period, starting in early 2019, with data focused on four U.S. cities: Boston, Dallas, Los Angeles, and Seattle. The researchers recorded visits to 433,000 specific “point of interest” locations in those cities, corroborated in part with records from Infogroup’s U.S. Business Database, an annual census of company information.  

      The researchers used U.S. Census Bureau data to categorize the socioeconomic status of the people in the study, placing everyone into one of four income quartiles, based on the average income of the census block (a small area) in which they live. The scholars made the same income-level assessment for every census block in the four cities, then recorded instances in which someone spent from 10 minutes to four hours in a census block other than their own, to see how often people visited areas in different income quartiles. 

      Ultimately, the researchers found that by late 2021, the amount of urban movement overall was returning to prepandemic levels, but the scope of places residents were visiting had become more restricted.

      Among other things, people made many fewer visits to museums, leisure venues, transport sites, and coffee shops. Visits to grocery stores remained fairly constant — but people tend not to leave their socioeconomic circles for grocery shopping.

      “Early in the pandemic, people reduced their mobility radius significantly,” Yabe says. “By late 2021, that decrease flattened out, and the average dwell time people spent at places other than work and home recovered to prepandemic levels. What’s different is that exploration substantially decreased, around 5 to 10 percent. We also see less visitation to fun places.” He adds: “Museums are the most diverse places you can find, parks — they took the biggest hit during the pandemic. Places that are [more] segregated, like grocery stores, did not.”

      Overall, Moro notes, “When we explore less, we go to places that are less diverse.”

      Different cities, same pattern

      Because the study encompassed four cities with different types of policies about reopening public sites and businesses during the pandemic, the researchers could also evaluate what impact public health policies had on urban movement. But even in these different settings, the same phenomenon emerged, with a narrower range of mobility occurring by late 2021.

      “Despite the substantial differences in how cities dealt with Covid-19, the decrease in diversity and the behavioral changes were surprisingly similar across the four cities,” Yabe observes.

      The researchers emphasize that these changes in urban movement can have long-term societal effects. Prior research has shown a significant association between a diversity of social connections and greater economic success for people in lower-income groups. And while some interactions between people in different income quartiles might be brief and transactional, the evidence suggests that, on aggregate, other more substantial connections have also been reduced. Additionally, the scholars note, the narrowing of experience can also weaken civic ties and valuable political connections.

      “It’s creating an urban fabric that is actually more brittle, in the sense that we are less exposed to other people,” Moro says. “We don’t get to know other people in the city, and that is very important for policies and public opinion. We need to convince people that new policies and laws would be fair. And the only way to do that is to know other people’s needs. If we don’t see them around the city, that will be impossible.”

      At the same time, Yabe adds, “I think there is a lot we can do from a policy standpoint to bring people back to places that used to be a lot more diverse.” The researchers are currently developing further studies related to cultural and public institutions, as well and transportation issues, to try to evaluate urban connectivity in additional detail.

      “The quantity of our mobility has recovered,” Yabe says. “The quality has really changed, and we’re more segregated as a result.” More

    • 150 Shares159 Views

      in Data Management & Statistics

      Martin Wainwright named director of the Institute for Data, Systems, and Society

      by

      Martin Wainwright, the Cecil H. Green Professor in MIT’s departments of Electrical Engineering and Computer Science (EECS) and Mathematics, has been named the new director of the Institute for Data, Systems, and Society (IDSS), effective July 1.

      “Martin is a widely recognized leader in statistics and machine learning — both in research and in education. In taking on this leadership role in the college, Martin will work to build up the human and institutional behavior component of IDSS, while strengthening initiatives in both policy and statistics, and collaborations within the institute, across MIT, and beyond,” says Daniel Huttenlocher, dean of the MIT Schwarzman College of Computing and the Henry Ellis Warren Professor of Electrical Engineering and Computer Science. “I look forward to working with him and supporting his efforts in this next chapter for IDSS.”

      “Martin holds a strong belief in the value of theoretical, experimental, and computational approaches to research and in facilitating connections between them. He also places much importance in having practical, as well as academic, impact,” says Asu Ozdaglar, deputy dean of academics for the MIT Schwarzman College of Computing, department head of EECS, and the MathWorks Professor of Electrical Engineering and Computer Science. “As the new director of IDSS, he will undoubtedly bring these tenets to the role in advancing the mission of IDSS and helping to shape its future.”

      A principal investigator in the Laboratory for Information and Decision Systems and the Statistics and Data Science Center, Wainwright joined the MIT faculty in July 2022 from the University of California at Berkeley, where he held the Howard Friesen Chair with a joint appointment between the departments of Electrical Engineering and Computer Science and Statistics.

      Wainwright received his bachelor’s degree in mathematics from the University of Waterloo, Canada, and doctoral degree in electrical engineering and computer science from MIT. He has received a number of awards and recognition, including an Alfred P. Sloan Foundation Fellowship, and best paper awards from the IEEE Signal Processing Society, IEEE Communications Society, and IEEE Information Theory and Communication Societies. He has also been honored with the Medallion Lectureship and Award from the Institute of Mathematical Statistics, and the COPSS Presidents’ Award from the Joint Statistical Societies. He was a section lecturer with the International Congress of Mathematicians in 2014 and received the Blackwell Award from the Institute of Mathematical Statistics in 2017.

      He is the author of “High-dimensional Statistics: A Non-Asymptotic Viewpoint” (Cambridge University Press, 2019), and is coauthor on several books, including on graphical models and on sparse statistical modeling.

      Wainwright succeeds Munther Dahleh, the William A. Coolidge Professor in EECS, who has helmed IDSS since its founding in 2015.

      “I am grateful to Munther and thank him for his leadership of IDSS. As the founding director, he has led the creation of a remarkable new part of MIT,” says Huttenlocher. More

    • 100 Shares199 Views

      in Data Management & Statistics

      Drones navigate unseen environments with liquid neural networks

      by

      In the vast, expansive skies where birds once ruled supreme, a new crop of aviators is taking flight. These pioneers of the air are not living creatures, but rather a product of deliberate innovation: drones. But these aren’t your typical flying bots, humming around like mechanical bees. Rather, they’re avian-inspired marvels that soar through the sky, guided by liquid neural networks to navigate ever-changing and unseen environments with precision and ease.

      Inspired by the adaptable nature of organic brains, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have introduced a method for robust flight navigation agents to master vision-based fly-to-target tasks in intricate, unfamiliar environments. The liquid neural networks, which can continuously adapt to new data inputs, showed prowess in making reliable decisions in unknown domains like forests, urban landscapes, and environments with added noise, rotation, and occlusion. These adaptable models, which outperformed many state-of-the-art counterparts in navigation tasks, could enable potential real-world drone applications like search and rescue, delivery, and wildlife monitoring.

      The researchers’ recent study, published today in Science Robotics, details how this new breed of agents can adapt to significant distribution shifts, a long-standing challenge in the field. The team’s new class of machine-learning algorithms, however, captures the causal structure of tasks from high-dimensional, unstructured data, such as pixel inputs from a drone-mounted camera. These networks can then extract crucial aspects of a task (i.e., understand the task at hand) and ignore irrelevant features, allowing acquired navigation skills to transfer targets seamlessly to new environments.

      Play video

      Drones navigate unseen environments with liquid neural networks.

      “We are thrilled by the immense potential of our learning-based control approach for robots, as it lays the groundwork for solving problems that arise when training in one environment and deploying in a completely distinct environment without additional training,” says Daniela Rus, CSAIL director and the Andrew (1956) and Erna Viterbi Professor of Electrical Engineering and Computer Science at MIT. “Our experiments demonstrate that we can effectively teach a drone to locate an object in a forest during summer, and then deploy the model in winter, with vastly different surroundings, or even in urban settings, with varied tasks such as seeking and following. This adaptability is made possible by the causal underpinnings of our solutions. These flexible algorithms could one day aid in decision-making based on data streams that change over time, such as medical diagnosis and autonomous driving applications.”

      A daunting challenge was at the forefront: Do machine-learning systems understand the task they are given from data when flying drones to an unlabeled object? And, would they be able to transfer their learned skill and task to new environments with drastic changes in scenery, such as flying from a forest to an urban landscape? What’s more, unlike the remarkable abilities of our biological brains, deep learning systems struggle with capturing causality, frequently over-fitting their training data and failing to adapt to new environments or changing conditions. This is especially troubling for resource-limited embedded systems, like aerial drones, that need to traverse varied environments and respond to obstacles instantaneously. 

      The liquid networks, in contrast, offer promising preliminary indications of their capacity to address this crucial weakness in deep learning systems. The team’s system was first trained on data collected by a human pilot, to see how they transferred learned navigation skills to new environments under drastic changes in scenery and conditions. Unlike traditional neural networks that only learn during the training phase, the liquid neural net’s parameters can change over time, making them not only interpretable, but more resilient to unexpected or noisy data. 

      In a series of quadrotor closed-loop control experiments, the drones underwent range tests, stress tests, target rotation and occlusion, hiking with adversaries, triangular loops between objects, and dynamic target tracking. They tracked moving targets, and executed multi-step loops between objects in never-before-seen environments, surpassing performance of other cutting-edge counterparts. 

      The team believes that the ability to learn from limited expert data and understand a given task while generalizing to new environments could make autonomous drone deployment more efficient, cost-effective, and reliable. Liquid neural networks, they noted, could enable autonomous air mobility drones to be used for environmental monitoring, package delivery, autonomous vehicles, and robotic assistants. 

      “The experimental setup presented in our work tests the reasoning capabilities of various deep learning systems in controlled and straightforward scenarios,” says MIT CSAIL Research Affiliate Ramin Hasani. “There is still so much room left for future research and development on more complex reasoning challenges for AI systems in autonomous navigation applications, which has to be tested before we can safely deploy them in our society.”

      “Robust learning and performance in out-of-distribution tasks and scenarios are some of the key problems that machine learning and autonomous robotic systems have to conquer to make further inroads in society-critical applications,” says Alessio Lomuscio, professor of AI safety in the Department of Computing at Imperial College London. “In this context, the performance of liquid neural networks, a novel brain-inspired paradigm developed by the authors at MIT, reported in this study is remarkable. If these results are confirmed in other experiments, the paradigm here developed will contribute to making AI and robotic systems more reliable, robust, and efficient.”

      Clearly, the sky is no longer the limit, but rather a vast playground for the boundless possibilities of these airborne marvels. 

      Hasani and PhD student Makram Chahine; Patrick Kao ’22, MEng ’22; and PhD student Aaron Ray SM ’21 wrote the paper with Ryan Shubert ’20, MEng ’22; MIT postdocs Mathias Lechner and Alexander Amini; and Rus.

      This research was supported, in part, by Schmidt Futures, the U.S. Air Force Research Laboratory, the U.S. Air Force Artificial Intelligence Accelerator, and the Boeing Co. More

    • 50 Shares159 Views

      in Data Management & Statistics

      Illuminating the money trail

      by

      You may not know this, but the U.S. imposes a 12.5 percent import tariff on imported flashlights. However, for a product category the federal government describes as “portable electric lamps designed to function by their own source of energy, other than flashlights,” the import tariff is just 3.5 percent.

      At a glance, this seems inexplicable. Why is one kind of self-powered portable light taxed more heavily than another? According to MIT political science professor In Song Kim, a policy discrepancy like this often stems from the difference in firms’ political power, as well as the extent to which firms are empowered by global production networks. This is a subject Kim has spent years examining in detail, producing original scholarly results while opening up a wealth of big data about politics to the public.

      “We all understand companies as being important economic agents,” Kim says. “But companies are political agents, too. They are very important political actors.”

      In particular, Kim’s work has illuminated the effects of lobbying upon U.S. trade policy. International trade is often presented as an unalloyed good, opening up markets and fueling growth. Beyond that, trade issues are usually described at the industry level; we hear about what the agriculture lobby or auto industry wants. But in reality, different firms want different things, even within the same industry.

      As Kim’s work shows, most firms lobby for policies pertaining to specific components of their products, and trade policy consists heavily of carve-outs for companies, not industry-wide standards. Firms making non-flashlight portable lights, it would seem, are good at lobbying, but the benefits clearly do not carry over to all portable light makers, as long as products are not perfect substitutes for each other. Meanwhile, as Kim’s research also shows, lobbying helps firms grow faster in size, even as lobbying-influenced policies may slow down the economy as a whole.

      “All our existing theories suggest that trade policy is a public good, in the sense that the benefits of open trade, the gains from trade, will be enjoyed by the public and will benefit the country as a whole,” Kim says. “But what I’ve learned is that trade policies are very, very granular. It’s become obvious to me that trade is no longer a public good. It’s actually a private good for individual companies.”

      Kim’s work includes over a dozen published journal articles over the last several years, several other forthcoming research papers, and a book he is currently writing. At the same time, Kim has created a public database, LobbyView, which tracks money in U.S. politics extending back to 1999. LobbyView, as an important collection of political information, has research, educational, and public-interest applications, enabling others, in academia or outside it, to further delve into the topic.

      “I want to contribute to the scholarly community, and I also want to create a public [resource] for our MIT community [and beyond], so we can all study politics through it,” Kim says.

      Keeping the public good in sight

      Kim grew up in South Korea, in a setting where politics was central to daily life. Kim’s grandfather, Kim jae-soon, was the Speaker of the National Assembly in South Korea from 1988 through 1990 and an important figure in the country’s government.

      “I’ve always been fascinated by politics,” says Kim, who remembers prominent political figures dropping by the family home when he was young. One of the principal lessons Kim learned about politics from his grandfather, however, was not about proximity to power, but the importance of public service. The enduring lesson of his family’s engagement with politics, Kim says, is that “I truly believe in contributing to the public good.”

      Kim’s found his own way of contributing to the public good not as a politician but as a scholar of politics. Kim received his BA in political science from Yonsei University in Seoul but decided he wanted to pursue graduate studies in the U.S. He earned an MA in law and diplomacy from the Fletcher School of Tufts University, then an MA in political science at George Washington University. By this time, Kim had become focused on the quantitative analysis of trade policy; for his PhD work, he attended Princeton University and was awarded his doctorate in 2014, joining the MIT faculty that year.

      Among the key pieces of research Kim has published, one paper, “Political Cleavages within Industry: Firm-level Lobbying for Trade Liberalization,” published in the American Political Science Review and growing out of his dissertation research, helped show how remarkably specialized many trade policies are. As of 2017, the U.S. had almost 17,000 types of products it made tariff decisions about. Many of these are the component parts of a product; about two-thirds of international trade consists of manufactured components that get shipped around during the production process, rather than raw goods or finished products. That paper won the 2018 Michael Wallerstein Award for the best published article in political economy in the previous year.

      Another 2017 paper Kim co-authored, “The Charmed Life of Superstar Exporters,” from the Journal of Politics, provides more empirical evidence of the differences among firms within an industry. The “superstar” firms that are the largest exporters tend to lobby the most about trade politics; a firm’s characteristics reveal more about its preferences for open trade than the possibility that its industry as a whole will gain a comparative advantage internationally.

      Kim often uses large-scale data and computational methods to study international trade and trade politics. Still another paper he has co-authored, “Measuring Trade Profile with Granular Product-level Trade Data,” published in the American Journal of Political Science in 2020, traces trade relationships in highly specific terms. Looking at over 2 billion observations of international trade data, Kim developed an algorithm to group countries based on which products they import and export. The methodology helps researchers to learn about the highly different developmental paths that countries follow, and about the deepening international competition between countries such as the U.S. and China.

      At other times, Kim has analyzed who is influencing trade policy. His paper “Mapping Political Communities,” from the journal Political Analysis in 2021, looks at the U.S. Congress and uses mandatory reports filed by lobbyists to build a picture of which interests groups are most closely connected to which politicians.

      Kim has published all his papers while balancing both his scholarly research and the public launch of LobbyView, which occurred in 2018. He was awarded tenure by MIT in the spring of 2022. Currently he is an associate professor in the Department of Political Science and a faculty affiliate of the Institute for Data, Systems, and Society.

      By the book

      Kim has continued to explore firm-level lobbying dynamics, although his recent research runs in a few directions. In a 2021 working paper, Kim and co-author Federico Huneeus of the Central Bank of Chile built a model estimating that eliminating lobbying in the U.S. could increase productivity by as much as 6 percent.

      “Political rents [favorable policies] given to particular companies might introduce inefficiencies or a misallocation of resources in the economy,” Kim says. “You could allocate those resources to more productive although politically inactive firms, but now they’re given to less productive and yet politically active big companies, increasing market concentration and monopolies.”

      Kim is on sabbatical during the 2022-23 academic year, working on a book about the importance of firms’ political activities in trade policymaking. The book will have an expansive timeframe, dating back to ancient times, which underscores the salience of trade policy across eras. At the same time, the book will analyze the distinctive features of modern trade politics with deepening global production networks.

      “I’m trying to allow people to learn about the history of trade politics, to show how the politics have changed over time,” Kim says. “In doing that, I’m also highlighting the importance of firm-to-firm trade and the emergence of new trade coalitions among firms in different countries and industries that are linked through the global production chain.”

      While continuing his own scholarly research, Kim still leads LobbyView, which he views both as a big data resource for any scholars interested in money in politics and an excellent teaching resource for his MIT classes, as students can tap into it for projects and papers. LobbyView contains so much data, in fact, that part of the challenge is finding ways to mine it effectively.

      “It really offers me an opportunity to work with MIT students,” Kim says of LobbyView. “What I think I can contribute is to bring those technologies to our understanding of politics. Having this unique data set can really allow students here to use technology to learn about politics, and I believe that fits the MIT identity.” More

    • 138 Shares129 Views

      in Data Management & Statistics

      A method for designing neural networks optimally suited for certain tasks

      by

      Neural networks, a type of machine-learning model, are being used to help humans complete a wide variety of tasks, from predicting if someone’s credit score is high enough to qualify for a loan to diagnosing whether a patient has a certain disease. But researchers still have only a limited understanding of how these models work. Whether a given model is optimal for certain task remains an open question.

      MIT researchers have found some answers. They conducted an analysis of neural networks and proved that they can be designed so they are “optimal,” meaning they minimize the probability of misclassifying borrowers or patients into the wrong category when the networks are given a lot of labeled training data. To achieve optimality, these networks must be built with a specific architecture.

      The researchers discovered that, in certain situations, the building blocks that enable a neural network to be optimal are not the ones developers use in practice. These optimal building blocks, derived through the new analysis, are unconventional and haven’t been considered before, the researchers say.

      In a paper published this week in the Proceedings of the National Academy of Sciences, they describe these optimal building blocks, called activation functions, and show how they can be used to design neural networks that achieve better performance on any dataset. The results hold even as the neural networks grow very large. This work could help developers select the correct activation function, enabling them to build neural networks that classify data more accurately in a wide range of application areas, explains senior author Caroline Uhler, a professor in the Department of Electrical Engineering and Computer Science (EECS).

      “While these are new activation functions that have never been used before, they are simple functions that someone could actually implement for a particular problem. This work really shows the importance of having theoretical proofs. If you go after a principled understanding of these models, that can actually lead you to new activation functions that you would otherwise never have thought of,” says Uhler, who is also co-director of the Eric and Wendy Schmidt Center at the Broad Institute of MIT and Harvard, and a researcher at MIT’s Laboratory for Information and Decision Systems (LIDS) and its Institute for Data, Systems and Society (IDSS).

      Joining Uhler on the paper are lead author Adityanarayanan Radhakrishnan, an EECS graduate student and an Eric and Wendy Schmidt Center Fellow, and Mikhail Belkin, a professor in the Halicioğlu Data Science Institute at the University of California at San Diego.

      Activation investigation

      A neural network is a type of machine-learning model that is loosely based on the human brain. Many layers of interconnected nodes, or neurons, process data. Researchers train a network to complete a task by showing it millions of examples from a dataset.

      For instance, a network that has been trained to classify images into categories, say dogs and cats, is given an image that has been encoded as numbers. The network performs a series of complex multiplication operations, layer by layer, until the result is just one number. If that number is positive, the network classifies the image a dog, and if it is negative, a cat.

      Activation functions help the network learn complex patterns in the input data. They do this by applying a transformation to the output of one layer before data are sent to the next layer. When researchers build a neural network, they select one activation function to use. They also choose the width of the network (how many neurons are in each layer) and the depth (how many layers are in the network.)

      “It turns out that, if you take the standard activation functions that people use in practice, and keep increasing the depth of the network, it gives you really terrible performance. We show that if you design with different activation functions, as you get more data, your network will get better and better,” says Radhakrishnan.

      He and his collaborators studied a situation in which a neural network is infinitely deep and wide — which means the network is built by continually adding more layers and more nodes — and is trained to perform classification tasks. In classification, the network learns to place data inputs into separate categories.

      “A clean picture”

      After conducting a detailed analysis, the researchers determined that there are only three ways this kind of network can learn to classify inputs. One method classifies an input based on the majority of inputs in the training data; if there are more dogs than cats, it will decide every new input is a dog. Another method classifies by choosing the label (dog or cat) of the training data point that most resembles the new input.

      The third method classifies a new input based on a weighted average of all the training data points that are similar to it. Their analysis shows that this is the only method of the three that leads to optimal performance. They identified a set of activation functions that always use this optimal classification method.

      “That was one of the most surprising things — no matter what you choose for an activation function, it is just going to be one of these three classifiers. We have formulas that will tell you explicitly which of these three it is going to be. It is a very clean picture,” he says.

      They tested this theory on a several classification benchmarking tasks and found that it led to improved performance in many cases. Neural network builders could use their formulas to select an activation function that yields improved classification performance, Radhakrishnan says.

      In the future, the researchers want to use what they’ve learned to analyze situations where they have a limited amount of data and for networks that are not infinitely wide or deep. They also want to apply this analysis to situations where data do not have labels.

      “In deep learning, we want to build theoretically grounded models so we can reliably deploy them in some mission-critical setting. This is a promising approach at getting toward something like that — building architectures in a theoretically grounded way that translates into better results in practice,” he says.

      This work was supported, in part, by the National Science Foundation, Office of Naval Research, the MIT-IBM Watson AI Lab, the Eric and Wendy Schmidt Center at the Broad Institute, and a Simons Investigator Award. More

    • 63 Shares179 Views

      in Data Management & Statistics

      Strengthening trust in machine-learning models

      by

      Probabilistic machine learning methods are becoming increasingly powerful tools in data analysis, informing a range of critical decisions across disciplines and applications, from forecasting election results to predicting the impact of microloans on addressing poverty.

      This class of methods uses sophisticated concepts from probability theory to handle uncertainty in decision-making. But the math is only one piece of the puzzle in determining their accuracy and effectiveness. In a typical data analysis, researchers make many subjective choices, or potentially introduce human error, that must also be assessed in order to cultivate users’ trust in the quality of decisions based on these methods.

      To address this issue, MIT computer scientist Tamara Broderick, associate professor in the Department of Electrical Engineering and Computer Science (EECS) and a member of the Laboratory for Information and Decision Systems (LIDS), and a team of researchers have developed a classification system — a “taxonomy of trust” — that defines where trust might break down in a data analysis and identifies strategies to strengthen trust at each step. The other researchers on the project are Professor Anna Smith at the University of Kentucky, professors Tian Zheng and Andrew Gelman at Columbia University, and Professor Rachael Meager at the London School of Economics. The team’s hope is to highlight concerns that are already well-studied and those that need more attention.

      In their paper, published in February in Science Advances, the researchers begin by detailing the steps in the data analysis process where trust might break down: Analysts make choices about what data to collect and which models, or mathematical representations, most closely mirror the real-life problem or question they are aiming to answer. They select algorithms to fit the model and use code to run those algorithms. Each of these steps poses unique challenges around building trust. Some components can be checked for accuracy in measurable ways. “Does my code have bugs?”, for example, is a question that can be tested against objective criteria. Other times, problems are more subjective, with no clear-cut answers; analysts are confronted with numerous strategies to gather data and decide whether a model reflects the real world.

      “What I think is nice about making this taxonomy, is that it really highlights where people are focusing. I think a lot of research naturally focuses on this level of ‘are my algorithms solving a particular mathematical problem?’ in part because it’s very objective, even if it’s a hard problem,” Broderick says.

      “I think it’s really hard to answer ‘is it reasonable to mathematize an important applied problem in a certain way?’ because it’s somehow getting into a harder space, it’s not just a mathematical problem anymore.”

      Capturing real life in a model

      The researchers’ work in categorizing where trust breaks down, though it may seem abstract, is rooted in real-world application.

      Meager, a co-author on the paper, analyzed whether microfinances can have a positive effect in a community. The project became a case study for where trust could break down, and ways to reduce this risk.

      At first look, measuring the impact of microfinancing might seem like a straightforward endeavor. But like any analysis, researchers meet challenges at each step in the process that can affect trust in the outcome. Microfinancing — in which individuals or small businesses receive small loans and other financial services in lieu of conventional banking — can offer different services, depending on the program. For the analysis, Meager gathered datasets from microfinance programs in countries across the globe, including in Mexico, Mongolia, Bosnia, and the Philippines.

      When combining conspicuously distinct datasets, in this case from multiple countries and across different cultures and geographies, researchers must evaluate whether specific case studies can reflect broader trends. It is also important to contextualize the data on hand. For example, in rural Mexico, owning goats may be counted as an investment.

      “It’s hard to measure the quality of life of an individual. People measure things like, ‘What’s the business profit of the small business?’ Or ‘What’s the consumption level of a household?’ There’s this potential for mismatch between what you ultimately really care about, and what you’re measuring,” Broderick says. “Before we get to the mathematical level, what data and what assumptions are we leaning on?”

      With data on hand, analysts must define the real-world questions they seek to answer. In the case of evaluating the benefits of microfinancing, analysts must define what they consider a positive outcome. It is standard in economics, for example, to measure the average financial gain per business in communities where a microfinance program is introduced. But reporting an average might suggest a net positive effect even if only a few (or even one) person benefited, instead of the community as a whole.

      “What you really wanted was that a lot of people are benefiting,” Broderick says. “It sounds simple. Why didn’t we measure the thing that we cared about? But I think it’s really common that practitioners use standard machine learning tools, for a lot of reasons. And these tools might report a proxy that doesn’t always agree with the quantity of interest.”

      Analysts may consciously or subconsciously favor models they are familiar with, especially after investing a great deal of time learning their ins and outs. “Someone might be hesitant to try a nonstandard method because they might be less certain they will use it correctly. Or peer review might favor certain familiar methods, even if a researcher might like to use nonstandard methods,” Broderick says. “There are a lot of reasons, sociologically. But this can be a concern for trust.”

      Final step, checking the code 

      While distilling a real-life problem into a model can be a big-picture, amorphous problem, checking the code that runs an algorithm can feel “prosaic,” Broderick says. But it is another potentially overlooked area where trust can be strengthened.

      In some cases, checking a coding pipeline that executes an algorithm might be considered outside the purview of an analyst’s job, especially when there is the option to use standard software packages.

      One way to catch bugs is to test whether code is reproducible. Depending on the field, however, sharing code alongside published work is not always a requirement or the norm. As models increase in complexity over time, it becomes harder to recreate code from scratch. Reproducing a model becomes difficult or even impossible.

      “Let’s just start with every journal requiring you to release your code. Maybe it doesn’t get totally double-checked, and everything isn’t absolutely perfect, but let’s start there,” Broderick says, as one step toward building trust.

      Paper co-author Gelman worked on an analysis that forecast the 2020 U.S. presidential election using state and national polls in real-time. The team published daily updates in The Economist magazine, while also publishing their code online for anyone to download and run themselves. Throughout the season, outsiders pointed out both bugs and conceptual problems in the model, ultimately contributing to a stronger analysis.

      The researchers acknowledge that while there is no single solution to create a perfect model, analysts and scientists have the opportunity to reinforce trust at nearly every turn.

      “I don’t think we expect any of these things to be perfect,” Broderick says, “but I think we can expect them to be better or to be as good as possible.” More