Machine learning

Subterms

    More stories

    • 150 Shares169 Views

      in Data Management & Statistics

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

      by

      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

    • 88 Shares179 Views

      in Data Management & Statistics

      The curse of variety in transportation systems

      by

      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

    • 113 Shares159 Views

      in Data Management & Statistics

      A simpler method for learning to control a robot

      by

      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

    • 88 Shares109 Views

      in Data Management & Statistics

      A faster way to teach a robot

      by

      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

    • 100 Shares189 Views

      in Data Management & Statistics

      A new way to look at data privacy

      by

      Imagine that a team of scientists has developed a machine-learning model that can predict whether a patient has cancer from lung scan images. They want to share this model with hospitals around the world so clinicians can start using it in diagnosis.

      But there’s a problem. To teach their model how to predict cancer, they showed it millions of real lung scan images, a process called training. Those sensitive data, which are now encoded into the inner workings of the model, could potentially be extracted by a malicious agent. The scientists can prevent this by adding noise, or more generic randomness, to the model that makes it harder for an adversary to guess the original data. However, perturbation reduces a model’s accuracy, so the less noise one can add, the better.

      MIT researchers have developed a technique that enables the user to potentially add the smallest amount of noise possible, while still ensuring the sensitive data are protected.

      The researchers created a new privacy metric, which they call Probably Approximately Correct (PAC) Privacy, and built a framework based on this metric that can automatically determine the minimal amount of noise that needs to be added. Moreover, this framework does not need knowledge of the inner workings of a model or its training process, which makes it easier to use for different types of models and applications.

      In several cases, the researchers show that the amount of noise required to protect sensitive data from adversaries is far less with PAC Privacy than with other approaches. This could help engineers create machine-learning models that provably hide training data, while maintaining accuracy in real-world settings.

      “PAC Privacy exploits the uncertainty or entropy of the sensitive data in a meaningful way,  and this allows us to add, in many cases, an order of magnitude less noise. This framework allows us to understand the characteristics of arbitrary data processing and privatize it automatically without artificial modifications. While we are in the early days and we are doing simple examples, we are excited about the promise of this technique,” says Srini Devadas, the Edwin Sibley Webster Professor of Electrical Engineering and co-author of a new paper on PAC Privacy.

      Devadas wrote the paper with lead author Hanshen Xiao, an electrical engineering and computer science graduate student. The research will be presented at the International Cryptography Conference (Crypto 2023).

      Defining privacy

      A fundamental question in data privacy is: How much sensitive data could an adversary recover from a machine-learning model with noise added to it?

      Differential Privacy, one popular privacy definition, says privacy is achieved if an adversary who observes the released model cannot infer whether an arbitrary individual’s data is used for the training processing. But provably preventing an adversary from distinguishing data usage often requires large amounts of noise to obscure it. This noise reduces the model’s accuracy.

      PAC Privacy looks at the problem a bit differently. It characterizes how hard it would be for an adversary to reconstruct any part of randomly sampled or generated sensitive data after noise has been added, rather than only focusing on the distinguishability problem.

      For instance, if the sensitive data are images of human faces, differential privacy would focus on whether the adversary can tell if someone’s face was in the dataset. PAC Privacy, on the other hand, could look at whether an adversary could extract a silhouette — an approximation — that someone could recognize as a particular individual’s face.

      Once they established the definition of PAC Privacy, the researchers created an algorithm that automatically tells the user how much noise to add to a model to prevent an adversary from confidently reconstructing a close approximation of the sensitive data. This algorithm guarantees privacy even if the adversary has infinite computing power, Xiao says.

      To find the optimal amount of noise, the PAC Privacy algorithm relies on the uncertainty, or entropy, in the original data from the viewpoint of the adversary.

      This automatic technique takes samples randomly from a data distribution or a large data pool and runs the user’s machine-learning training algorithm on that subsampled data to produce an output learned model. It does this many times on different subsamplings and compares the variance across all outputs. This variance determines how much noise one must add — a smaller variance means less noise is needed.

      Algorithm advantages

      Different from other privacy approaches, the PAC Privacy algorithm does not need knowledge of the inner workings of a model, or the training process.

      When implementing PAC Privacy, a user can specify their desired level of confidence at the outset. For instance, perhaps the user wants a guarantee that an adversary will not be more than 1 percent confident that they have successfully reconstructed the sensitive data to within 5 percent of its actual value. The PAC Privacy algorithm automatically tells the user the optimal amount of noise that needs to be added to the output model before it is shared publicly, in order to achieve those goals.

      “The noise is optimal, in the sense that if you add less than we tell you, all bets could be off. But the effect of adding noise to neural network parameters is complicated, and we are making no promises on the utility drop the model may experience with the added noise,” Xiao says.

      This points to one limitation of PAC Privacy — the technique does not tell the user how much accuracy the model will lose once the noise is added. PAC Privacy also involves repeatedly training a machine-learning model on many subsamplings of data, so it can be computationally expensive.  

      To improve PAC Privacy, one approach is to modify a user’s machine-learning training process so it is more stable, meaning that the output model it produces does not change very much when the input data is subsampled from a data pool.  This stability would create smaller variances between subsample outputs, so not only would the PAC Privacy algorithm need to be run fewer times to identify the optimal amount of noise, but it would also need to add less noise.

      An added benefit of stabler models is that they often have less generalization error, which means they can make more accurate predictions on previously unseen data, a win-win situation between machine learning and privacy, Devadas adds.

      “In the next few years, we would love to look a little deeper into this relationship between stability and privacy, and the relationship between privacy and generalization error. We are knocking on a door here, but it is not clear yet where the door leads,” he says.

      “Obfuscating the usage of an individual’s data in a model is paramount to protecting their privacy. However, to do so can come at the cost of the datas’ and therefore model’s utility,” says Jeremy Goodsitt, senior machine learning engineer at Capital One, who was not involved with this research. “PAC provides an empirical, black-box solution, which can reduce the added noise compared to current practices while maintaining equivalent privacy guarantees. In addition, its empirical approach broadens its reach to more data consuming applications.”

      This research is funded, in part, by DSTA Singapore, Cisco Systems, Capital One, and a MathWorks Fellowship. More

    • 88 Shares159 Views

      in Data Management & Statistics

      3 Questions: Honing robot perception and mapping

      by

      Walking to a friend’s house or browsing the aisles of a grocery store might feel like simple tasks, but they in fact require sophisticated capabilities. That’s because humans are able to effortlessly understand their surroundings and detect complex information about patterns, objects, and their own location in the environment.

      What if robots could perceive their environment in a similar way? That question is on the minds of MIT Laboratory for Information and Decision Systems (LIDS) researchers Luca Carlone and Jonathan How. In 2020, a team led by Carlone released the first iteration of Kimera, an open-source library that enables a single robot to construct a three-dimensional map of its environment in real time, while labeling different objects in view. Last year, Carlone’s and How’s research groups (SPARK Lab and Aerospace Controls Lab) introduced Kimera-Multi, an updated system in which multiple robots communicate among themselves in order to create a unified map. A 2022 paper associated with the project recently received this year’s IEEE Transactions on Robotics King-Sun Fu Memorial Best Paper Award, given to the best paper published in the journal in 2022.

      Carlone, who is the Leonardo Career Development Associate Professor of Aeronautics and Astronautics, and How, the Richard Cockburn Maclaurin Professor in Aeronautics and Astronautics, spoke to LIDS about Kimera-Multi and the future of how robots might perceive and interact with their environment.

      Q: Currently your labs are focused on increasing the number of robots that can work together in order to generate 3D maps of the environment. What are some potential advantages to scaling this system?

      How: The key benefit hinges on consistency, in the sense that a robot can create an independent map, and that map is self-consistent but not globally consistent. We’re aiming for the team to have a consistent map of the world; that’s the key difference in trying to form a consensus between robots as opposed to mapping independently.

      Carlone: In many scenarios it’s also good to have a bit of redundancy. For example, if we deploy a single robot in a search-and-rescue mission, and something happens to that robot, it would fail to find the survivors. If multiple robots are doing the exploring, there’s a much better chance of success. Scaling up the team of robots also means that any given task may be completed in a shorter amount of time.

      Q: What are some of the lessons you’ve learned from recent experiments, and challenges you’ve had to overcome while designing these systems?

      Carlone: Recently we did a big mapping experiment on the MIT campus, in which eight robots traversed up to 8 kilometers in total. The robots have no prior knowledge of the campus, and no GPS. Their main tasks are to estimate their own trajectory and build a map around it. You want the robots to understand the environment as humans do; humans not only understand the shape of obstacles, to get around them without hitting them, but also understand that an object is a chair, a desk, and so on. There’s the semantics part.

      The interesting thing is that when the robots meet each other, they exchange information to improve their map of the environment. For instance, if robots connect, they can leverage information to correct their own trajectory. The challenge is that if you want to reach a consensus between robots, you don’t have the bandwidth to exchange too much data. One of the key contributions of our 2022 paper is to deploy a distributed protocol, in which robots exchange limited information but can still agree on how the map looks. They don’t send camera images back and forth but only exchange specific 3D coordinates and clues extracted from the sensor data. As they continue to exchange such data, they can form a consensus.

      Right now we are building color-coded 3D meshes or maps, in which the color contains some semantic information, like “green” corresponds to grass, and “magenta” to a building. But as humans, we have a much more sophisticated understanding of reality, and we have a lot of prior knowledge about relationships between objects. For instance, if I was looking for a bed, I would go to the bedroom instead of exploring the entire house. If you start to understand the complex relationships between things, you can be much smarter about what the robot can do in the environment. We’re trying to move from capturing just one layer of semantics, to a more hierarchical representation in which the robots understand rooms, buildings, and other concepts.

      Q: What kinds of applications might Kimera and similar technologies lead to in the future?

      How: Autonomous vehicle companies are doing a lot of mapping of the world and learning from the environments they’re in. The holy grail would be if these vehicles could communicate with each other and share information, then they could improve models and maps that much quicker. The current solutions out there are individualized. If a truck pulls up next to you, you can’t see in a certain direction. Could another vehicle provide a field of view that your vehicle otherwise doesn’t have? This is a futuristic idea because it requires vehicles to communicate in new ways, and there are privacy issues to overcome. But if we could resolve those issues, you could imagine a significantly improved safety situation, where you have access to data from multiple perspectives, not only your field of view.

      Carlone: These technologies will have a lot of applications. Earlier I mentioned search and rescue. Imagine that you want to explore a forest and look for survivors, or map buildings after an earthquake in a way that can help first responders access people who are trapped. Another setting where these technologies could be applied is in factories. Currently, robots that are deployed in factories are very rigid. They follow patterns on the floor, and are not really able to understand their surroundings. But if you’re thinking about much more flexible factories in the future, robots will have to cooperate with humans and exist in a much less structured environment. More

    • 113 Shares189 Views

      in Data Management & Statistics

      Learning the language of molecules to predict their properties

      by

      Discovering new materials and drugs typically involves a manual, trial-and-error process that can take decades and cost millions of dollars. To streamline this process, scientists often use machine learning to predict molecular properties and narrow down the molecules they need to synthesize and test in the lab.

      Researchers from MIT and the MIT-Watson AI Lab have developed a new, unified framework that can simultaneously predict molecular properties and generate new molecules much more efficiently than these popular deep-learning approaches.

      To teach a machine-learning model to predict a molecule’s biological or mechanical properties, researchers must show it millions of labeled molecular structures — a process known as training. Due to the expense of discovering molecules and the challenges of hand-labeling millions of structures, large training datasets are often hard to come by, which limits the effectiveness of machine-learning approaches.

      By contrast, the system created by the MIT researchers can effectively predict molecular properties using only a small amount of data. Their system has an underlying understanding of the rules that dictate how building blocks combine to produce valid molecules. These rules capture the similarities between molecular structures, which helps the system generate new molecules and predict their properties in a data-efficient manner.

      This method outperformed other machine-learning approaches on both small and large datasets, and was able to accurately predict molecular properties and generate viable molecules when given a dataset with fewer than 100 samples.

      “Our goal with this project is to use some data-driven methods to speed up the discovery of new molecules, so you can train a model to do the prediction without all of these cost-heavy experiments,” says lead author Minghao Guo, a computer science and electrical engineering (EECS) graduate student.

      Guo’s co-authors include MIT-IBM Watson AI Lab research staff members Veronika Thost, Payel Das, and Jie Chen; recent MIT graduates Samuel Song ’23 and Adithya Balachandran ’23; and senior author Wojciech Matusik, a professor of electrical engineering and computer science and a member of the MIT-IBM Watson AI Lab, who leads the Computational Design and Fabrication Group within the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). The research will be presented at the International Conference for Machine Learning.

      Learning the language of molecules

      To achieve the best results with machine-learning models, scientists need training datasets with millions of molecules that have similar properties to those they hope to discover. In reality, these domain-specific datasets are usually very small. So, researchers use models that have been pretrained on large datasets of general molecules, which they apply to a much smaller, targeted dataset. However, because these models haven’t acquired much domain-specific knowledge, they tend to perform poorly.

      The MIT team took a different approach. They created a machine-learning system that automatically learns the “language” of molecules — what is known as a molecular grammar — using only a small, domain-specific dataset. It uses this grammar to construct viable molecules and predict their properties.

      In language theory, one generates words, sentences, or paragraphs based on a set of grammar rules. You can think of a molecular grammar the same way. It is a set of production rules that dictate how to generate molecules or polymers by combining atoms and substructures.

      Just like a language grammar, which can generate a plethora of sentences using the same rules, one molecular grammar can represent a vast number of molecules. Molecules with similar structures use the same grammar production rules, and the system learns to understand these similarities.

      Since structurally similar molecules often have similar properties, the system uses its underlying knowledge of molecular similarity to predict properties of new molecules more efficiently. 

      “Once we have this grammar as a representation for all the different molecules, we can use it to boost the process of property prediction,” Guo says.

      The system learns the production rules for a molecular grammar using reinforcement learning — a trial-and-error process where the model is rewarded for behavior that gets it closer to achieving a goal.

      But because there could be billions of ways to combine atoms and substructures, the process to learn grammar production rules would be too computationally expensive for anything but the tiniest dataset.

      The researchers decoupled the molecular grammar into two parts. The first part, called a metagrammar, is a general, widely applicable grammar they design manually and give the system at the outset. Then it only needs to learn a much smaller, molecule-specific grammar from the domain dataset. This hierarchical approach speeds up the learning process.

      Big results, small datasets

      In experiments, the researchers’ new system simultaneously generated viable molecules and polymers, and predicted their properties more accurately than several popular machine-learning approaches, even when the domain-specific datasets had only a few hundred samples. Some other methods also required a costly pretraining step that the new system avoids.

      The technique was especially effective at predicting physical properties of polymers, such as the glass transition temperature, which is the temperature required for a material to transition from solid to liquid. Obtaining this information manually is often extremely costly because the experiments require extremely high temperatures and pressures.

      To push their approach further, the researchers cut one training set down by more than half — to just 94 samples. Their model still achieved results that were on par with methods trained using the entire dataset.

      “This grammar-based representation is very powerful. And because the grammar itself is a very general representation, it can be deployed to different kinds of graph-form data. We are trying to identify other applications beyond chemistry or material science,” Guo says.

      In the future, they also want to extend their current molecular grammar to include the 3D geometry of molecules and polymers, which is key to understanding the interactions between polymer chains. They are also developing an interface that would show a user the learned grammar production rules and solicit feedback to correct rules that may be wrong, boosting the accuracy of the system.

      This work is funded, in part, by the MIT-IBM Watson AI Lab and its member company, Evonik. More

    • 88 Shares189 Views

      in Data Management & Statistics

      Educating national security leaders on artificial intelligence

      by

      Understanding artificial intelligence and how it relates to matters of national security has become a top priority for military and government leaders in recent years. A new three-day custom program entitled “Artificial Intelligence for National Security Leaders” — AI4NSL for short — aims to educate leaders who may not have a technical background on the basics of AI, machine learning, and data science, and how these topics intersect with national security.

      “National security fundamentally is about two things: getting information out of sensors and processing that information. These are two things that AI excels at. The AI4NSL class engages national security leaders in understanding how to navigate the benefits and opportunities that AI affords, while also understanding its potential negative consequences,” says Aleksander Madry, the Cadence Design Systems Professor at MIT and one of the course’s faculty directors.

      Organized jointly by MIT’s School of Engineering, MIT Stephen A. Schwarzman College of Computing, and MIT Sloan Executive Education, AI4NSL wrapped up its fifth cohort in April. The course brings leaders from every branch of the U.S. military, as well as some foreign military leaders from NATO, to MIT’s campus, where they learn from faculty experts on a variety of technical topics in AI, as well as how to navigate organizational challenges that arise in this context.

      Play video

      AI for National Security Leaders | MIT Sloan Executive Education

      “We set out to put together a real executive education class on AI for senior national security leaders,” says Madry. “For three days, we are teaching these leaders not only an understanding of what this technology is about, but also how to best adopt these technologies organizationally.”

      The original idea sprang from discussions with senior U.S. Air Force (USAF) leaders and members of the Department of the Air Force (DAF)-MIT AI Accelerator in 2019.

      According to Major John Radovan, deputy director of the DAF-MIT AI Accelerator, in recent years it has become clear that national security leaders needed a deeper understanding of AI technologies and its implications on security, warfare, and military operations. In February 2020, Radovan and his team at the DAF-MIT AI Accelerator started building a custom course to help guide senior leaders in their discussions about AI.

      “This is the only course out there that is focused on AI specifically for national security,” says Radovan. “We didn’t want to make this course just for members of the Air Force — it had to be for all branches of the military. If we are going to operate as a joint force, we need to have the same vocabulary and the same mental models about how to use this technology.”

      After a pilot program in collaboration with MIT Open Learning and the MIT Computer Science and Artificial Intelligence Laboratory, Radovan connected with faculty at the School of Engineering and MIT Schwarzman College of Computing, including Madry, to refine the course’s curriculum. They enlisted the help of colleagues and faculty at MIT Sloan Executive Education to refine the class’s curriculum and cater the content to its audience. The result of this cross-school collaboration was a new iteration of AI4NSL, which was launched last summer.

      In addition to providing participants with a basic overview of AI technologies, the course places a heavy emphasis on organizational planning and implementation.

      “What we wanted to do was to create smart consumers at the command level. The idea was to present this content at a higher level so that people could understand the key frameworks, which will guide their thinking around the use and adoption of this material,” says Roberto Fernandez, the William F. Pounds Professor of Management and one of the AI4NSL instructors, as well as the other course’s faculty director.

      During the three-day course, instructors from MIT’s Department of Electrical Engineering and Computer Science, Department of Aeronautics and Astronautics, and MIT Sloan School of Management cover a wide range of topics.

      The first half of the course starts with a basic overview of concepts including AI, machine learning, deep learning, and the role of data. Instructors also present the problems and pitfalls of using AI technologies, including the potential for adversarial manipulation of machine learning systems, privacy challenges, and ethical considerations.

      In the middle of day two, the course shifts to examine the organizational perspective, encouraging participants to consider how to effectively implement these technologies in their own units.

      “What’s exciting about this course is the way it is formatted first in terms of understanding AI, machine learning, what data is, and how data feeds AI, and then giving participants a framework to go back to their units and build a strategy to make this work,” says Colonel Michelle Goyette, director of the Army Strategic Education Program at the Army War College and an AI4NSL participant.

      Throughout the course, breakout sessions provide participants with an opportunity to collaborate and problem-solve on an exercise together. These breakout sessions build upon one another as the participants are exposed to new concepts related to AI.

      “The breakout sessions have been distinctive because they force you to establish relationships with people you don’t know, so the networking aspect is key. Any time you can do more than receive information and actually get into the application of what you were taught, that really enhances the learning environment,” says Lieutenant General Brian Robinson, the commander of Air Education and Training Command for the USAF and an AI4NSL participant.

      This spirit of teamwork, collaboration, and bringing together individuals from different backgrounds permeates the three-day program. The AI4NSL classroom not only brings together national security leaders from all branches of the military, it also brings together faculty from three schools across MIT.

      “One of the things that’s most exciting about this program is the kind of overarching theme of collaboration,” says Rob Dietel, director of executive programs at Sloan School of Management. “We’re not drawing just from the MIT Sloan faculty, we’re bringing in top faculty from the Schwarzman College of Computing and the School of Engineering. It’s wonderful to be able to tap into those resources that are here on MIT’s campus to really make it the most impactful program that we can.”

      As new developments in generative AI, such as ChatGPT, and machine learning alter the national security landscape, the organizers at AI4NSL will continue to update the curriculum to ensure it is preparing leaders to understand the implications for their respective units.

      “The rate of change for AI and national security is so fast right now that it’s challenging to keep up, and that’s part of the reason we’ve designed this program. We’ve brought in some of our world-class faculty from different parts of MIT to really address the changing dynamic of AI,” adds Dietel. More