More stories

  • in

    Automated method helps researchers quantify uncertainty in their predictions

    Pollsters trying to predict presidential election results and physicists searching for distant exoplanets have at least one thing in common: They often use a tried-and-true scientific technique called Bayesian inference.

    Bayesian inference allows these scientists to effectively estimate some unknown parameter — like the winner of an election — from data such as poll results. But Bayesian inference can be slow, sometimes consuming weeks or even months of computation time or requiring a researcher to spend hours deriving tedious equations by hand. 

    Researchers from MIT and elsewhere have introduced an optimization technique that speeds things up without requiring a scientist to do a lot of additional work. Their method can achieve more accurate results faster than another popular approach for accelerating Bayesian inference.

    Using this new automated technique, a scientist could simply input their model and then the optimization method does all the calculations under the hood to provide an approximation of some unknown parameter. The method also offers reliable uncertainty estimates that can help a researcher understand when to trust its predictions.

    This versatile technique could be applied to a wide array of scientific quandaries that incorporate Bayesian inference. For instance, it could be used by economists studying the impact of microcredit loans in developing nations or sports analysts using a model to rank top tennis players.

    “When you actually dig into what people are doing in the social sciences, physics, chemistry, or biology, they are often using a lot of the same tools under the hood. There are so many Bayesian analyses out there. If we can build a really great tool that makes these researchers lives easier, then we can really make a difference to a lot of people in many different research areas,” says senior author Tamara Broderick, an associate professor in MIT’s Department of Electrical Engineering and Computer Science (EECS) and a member of the Laboratory for Information and Decision Systems and the Institute for Data, Systems, and Society.

    Broderick is joined on the paper by co-lead authors Ryan Giordano, an assistant professor of statistics at the University of California at Berkeley; and Martin Ingram, a data scientist at the AI company KONUX. The paper was recently published in the Journal of Machine Learning Research.

    Faster results

    When researchers seek a faster form of Bayesian inference, they often turn to a technique called automatic differentiation variational inference (ADVI), which is often both fast to run and easy to use.

    But Broderick and her collaborators have found a number of practical issues with ADVI. It has to solve an optimization problem and can do so only approximately. So, ADVI can still require a lot of computation time and user effort to determine whether the approximate solution is good enough. And once it arrives at a solution, it tends to provide poor uncertainty estimates.

    Rather than reinventing the wheel, the team took many ideas from ADVI but turned them around to create a technique called deterministic ADVI (DADVI) that doesn’t have these downsides.

    With DADVI, it is very clear when the optimization is finished, so a user won’t need to spend extra computation time to ensure that the best solution has been found. DADVI also permits the incorporation of more powerful optimization methods that give it an additional speed and performance boost.

    Once it reaches a result, DADVI is set up to allow the use of uncertainty corrections. These corrections make its uncertainty estimates much more accurate than those of ADVI.

    DADVI also enables the user to clearly see how much error they have incurred in the approximation to the optimization problem. This prevents a user from needlessly running the optimization again and again with more and more resources to try and reduce the error.

    “We wanted to see if we could live up to the promise of black-box inference in the sense of, once the user makes their model, they can just run Bayesian inference and don’t have to derive everything by hand, they don’t need to figure out when to stop their algorithm, and they have a sense of how accurate their approximate solution is,” Broderick says.

    Defying conventional wisdom

    DADVI can be more effective than ADVI because it uses an efficient approximation method, called sample average approximation, which estimates an unknown quantity by taking a series of exact steps.

    Because the steps along the way are exact, it is clear when the objective has been reached. Plus, getting to that objective typically requires fewer steps.

    Often, researchers expect sample average approximation to be more computationally intensive than a more popular method, known as stochastic gradient, which is used by ADVI. But Broderick and her collaborators showed that, in many applications, this is not the case.

    “A lot of problems really do have special structure, and you can be so much more efficient and get better performance by taking advantage of that special structure. That is something we have really seen in this paper,” she adds.

    They tested DADVI on a number of real-world models and datasets, including a model used by economists to evaluate the effectiveness of microcredit loans and one used in ecology to determine whether a species is present at a particular site.

    Across the board, they found that DADVI can estimate unknown parameters faster and more reliably than other methods, and achieves as good or better accuracy than ADVI. Because it is easier to use than other techniques, DADVI could offer a boost to scientists in a wide variety of fields.

    In the future, the researchers want to dig deeper into correction methods for uncertainty estimates so they can better understand why these corrections can produce such accurate uncertainties, and when they could fall short.

    “In applied statistics, we often have to use approximate algorithms for problems that are too complex or high-dimensional to allow exact solutions to be computed in reasonable time. This new paper offers an interesting set of theory and empirical results that point to an improvement in a popular existing approximate algorithm for Bayesian inference,” says Andrew Gelman ’85, ’86, a professor of statistics and political science at Columbia University, who was not involved with the study. “As one of the team involved in the creation of that earlier work, I’m happy to see our algorithm superseded by something more stable.”

    This research was supported by a National Science Foundation CAREER Award and the U.S. Office of Naval Research.  More

  • in

    Study: Global deforestation leads to more mercury pollution

    About 10 percent of human-made mercury emissions into the atmosphere each year are the result of global deforestation, according to a new MIT study.

    The world’s vegetation, from the Amazon rainforest to the savannahs of sub-Saharan Africa, acts as a sink that removes the toxic pollutant from the air. However, if the current rate of deforestation remains unchanged or accelerates, the researchers estimate that net mercury emissions will keep increasing.

    “We’ve been overlooking a significant source of mercury, especially in tropical regions,” says Ari Feinberg, a former postdoc in the Institute for Data, Systems, and Society (IDSS) and lead author of the study.

    The researchers’ model shows that the Amazon rainforest plays a particularly important role as a mercury sink, contributing about 30 percent of the global land sink. Curbing Amazon deforestation could thus have a substantial impact on reducing mercury pollution.

    The team also estimates that global reforestation efforts could increase annual mercury uptake by about 5 percent. While this is significant, the researchers emphasize that reforestation alone should not be a substitute for worldwide pollution control efforts.

    “Countries have put a lot of effort into reducing mercury emissions, especially northern industrialized countries, and for very good reason. But 10 percent of the global anthropogenic source is substantial, and there is a potential for that to be even greater in the future. [Addressing these deforestation-related emissions] needs to be part of the solution,” says senior author Noelle Selin, a professor in IDSS and MIT’s Department of Earth, Atmospheric and Planetary Sciences.

    Feinberg and Selin are joined on the paper by co-authors Martin Jiskra, a former Swiss National Science Foundation Ambizione Fellow at the University of Basel; Pasquale Borrelli, a professor at Roma Tre University in Italy; and Jagannath Biswakarma, a postdoc at the Swiss Federal Institute of Aquatic Science and Technology. The paper appears today in Environmental Science and Technology.

    Modeling mercury

    Over the past few decades, scientists have generally focused on studying deforestation as a source of global carbon dioxide emissions. Mercury, a trace element, hasn’t received the same attention, partly because the terrestrial biosphere’s role in the global mercury cycle has only recently been better quantified.

    Plant leaves take up mercury from the atmosphere, in a similar way as they take up carbon dioxide. But unlike carbon dioxide, mercury doesn’t play an essential biological function for plants. Mercury largely stays within a leaf until it falls to the forest floor, where the mercury is absorbed by the soil.

    Mercury becomes a serious concern for humans if it ends up in water bodies, where it can become methylated by microorganisms. Methylmercury, a potent neurotoxin, can be taken up by fish and bioaccumulated through the food chain. This can lead to risky levels of methylmercury in the fish humans eat.

    “In soils, mercury is much more tightly bound than it would be if it were deposited in the ocean. The forests are doing a sort of ecosystem service, in that they are sequestering mercury for longer timescales,” says Feinberg, who is now a postdoc in the Blas Cabrera Institute of Physical Chemistry in Spain.

    In this way, forests reduce the amount of toxic methylmercury in oceans.

    Many studies of mercury focus on industrial sources, like burning fossil fuels, small-scale gold mining, and metal smelting. A global treaty, the 2013 Minamata Convention, calls on nations to reduce human-made emissions. However, it doesn’t directly consider impacts of deforestation.

    The researchers launched their study to fill in that missing piece.

    In past work, they had built a model to probe the role vegetation plays in mercury uptake. Using a series of land use change scenarios, they adjusted the model to quantify the role of deforestation.

    Evaluating emissions

    This chemical transport model tracks mercury from its emissions sources to where it is chemically transformed in the atmosphere and then ultimately to where it is deposited, mainly through rainfall or uptake into forest ecosystems.

    They divided the Earth into eight regions and performed simulations to calculate deforestation emissions factors for each, considering elements like type and density of vegetation, mercury content in soils, and historical land use.

    However, good data for some regions were hard to come by.

    They lacked measurements from tropical Africa or Southeast Asia — two areas that experience heavy deforestation. To get around this gap, they used simpler, offline models to simulate hundreds of scenarios, which helped them improve their estimations of potential uncertainties.

    They also developed a new formulation for mercury emissions from soil. This formulation captures the fact that deforestation reduces leaf area, which increases the amount of sunlight that hits the ground and accelerates the outgassing of mercury from soils.

    The model divides the world into grid squares, each of which is a few hundred square kilometers. By changing land surface and vegetation parameters in certain squares to represent deforestation and reforestation scenarios, the researchers can capture impacts on the mercury cycle.

    Overall, they found that about 200 tons of mercury are emitted to the atmosphere as the result of deforestation, or about 10 percent of total human-made emissions. But in tropical and sub-tropical countries, deforestation emissions represent a higher percentage of total emissions. For example, in Brazil deforestation emissions are 40 percent of total human-made emissions.

    In addition, people often light fires to prepare tropical forested areas for agricultural activities, which causes more emissions by releasing mercury stored by vegetation.

    “If deforestation was a country, it would be the second highest emitting country, after China, which emits around 500 tons of mercury a year,” Feinberg adds.

    And since the Minamata Convention is now addressing primary mercury emissions, scientists can expect deforestation to become a larger fraction of human-made emissions in the future.

    “Policies to protect forests or cut them down have unintended effects beyond their target. It is important to consider the fact that these are systems, and they involve human activities, and we need to understand them better in order to actually solve the problems that we know are out there,” Selin says.

    By providing this first estimate, the team hopes to inspire more research in this area.

    In the future, they want to incorporate more dynamic Earth system models into their analysis, which would enable them to interactively track mercury uptake and better model the timescale of vegetation regrowth.

    “This paper represents an important advance in our understanding of global mercury cycling by quantifying a pathway that has long been suggested but not yet quantified. Much of our research to date has focused on primary anthropogenic emissions — those directly resulting from human activity via coal combustion or mercury-gold amalgam burning in artisanal and small-scale gold mining,” says Jackie Gerson, an assistant professor in the Department of Earth and Environmental Sciences at Michigan State University, who was not involved with this research. “This research shows that deforestation can also result in substantial mercury emissions and needs to be considered both in terms of global mercury models and land management policies. It therefore has the potential to advance our field scientifically as well as to promote policies that reduce mercury emissions via deforestation.

    This work was funded, in part, by the U.S. National Science Foundation, the Swiss National Science Foundation, and Swiss Federal Institute of Aquatic Science and Technology. More

  • in

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

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

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

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

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

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

    Prismatic perspectives

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

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

    The path forward

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

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

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

  • in

    Multiple AI models help robots execute complex plans more transparently

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

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

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

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

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

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

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

    A three-pronged hierarchy

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

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

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

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

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

  • in

    AI accelerates problem-solving in complex scenarios

    While Santa Claus may have a magical sleigh and nine plucky reindeer to help him deliver presents, for companies like FedEx, the optimization problem of efficiently routing holiday packages is so complicated that they often employ specialized software to find a solution.

    This software, called a mixed-integer linear programming (MILP) solver, splits a massive optimization problem into smaller pieces and uses generic algorithms to try and find the best solution. However, the solver could take hours — or even days — to arrive at a solution.

    The process is so onerous that a company often must stop the software partway through, accepting a solution that is not ideal but the best that could be generated in a set amount of time.

    Researchers from MIT and ETH Zurich used machine learning to speed things up.

    They identified a key intermediate step in MILP solvers that has so many potential solutions it takes an enormous amount of time to unravel, which slows the entire process. The researchers employed a filtering technique to simplify this step, then used machine learning to find the optimal solution for a specific type of problem.

    Their data-driven approach enables a company to use its own data to tailor a general-purpose MILP solver to the problem at hand.

    This new technique sped up MILP solvers between 30 and 70 percent, without any drop in accuracy. One could use this method to obtain an optimal solution more quickly or, for especially complex problems, a better solution in a tractable amount of time.

    This approach could be used wherever MILP solvers are employed, such as by ride-hailing services, electric grid operators, vaccination distributors, or any entity faced with a thorny resource-allocation problem.

    “Sometimes, in a field like optimization, it is very common for folks to think of solutions as either purely machine learning or purely classical. I am a firm believer that we want to get the best of both worlds, and this is a really strong instantiation of that hybrid approach,” says senior author Cathy Wu, the Gilbert W. Winslow Career Development Assistant Professor in Civil and Environmental Engineering (CEE), and a member of a member of the Laboratory for Information and Decision Systems (LIDS) and the Institute for Data, Systems, and Society (IDSS).

    Wu wrote the paper with co-lead authors Siriu Li, an IDSS graduate student, and Wenbin Ouyang, a CEE graduate student; as well as Max Paulus, a graduate student at ETH Zurich. The research will be presented at the Conference on Neural Information Processing Systems.

    Tough to solve

    MILP problems have an exponential number of potential solutions. For instance, say a traveling salesperson wants to find the shortest path to visit several cities and then return to their city of origin. If there are many cities which could be visited in any order, the number of potential solutions might be greater than the number of atoms in the universe.  

    “These problems are called NP-hard, which means it is very unlikely there is an efficient algorithm to solve them. When the problem is big enough, we can only hope to achieve some suboptimal performance,” Wu explains.

    An MILP solver employs an array of techniques and practical tricks that can achieve reasonable solutions in a tractable amount of time.

    A typical solver uses a divide-and-conquer approach, first splitting the space of potential solutions into smaller pieces with a technique called branching. Then, the solver employs a technique called cutting to tighten up these smaller pieces so they can be searched faster.

    Cutting uses a set of rules that tighten the search space without removing any feasible solutions. These rules are generated by a few dozen algorithms, known as separators, that have been created for different kinds of MILP problems. 

    Wu and her team found that the process of identifying the ideal combination of separator algorithms to use is, in itself, a problem with an exponential number of solutions.

    “Separator management is a core part of every solver, but this is an underappreciated aspect of the problem space. One of the contributions of this work is identifying the problem of separator management as a machine learning task to begin with,” she says.

    Shrinking the solution space

    She and her collaborators devised a filtering mechanism that reduces this separator search space from more than 130,000 potential combinations to around 20 options. This filtering mechanism draws on the principle of diminishing marginal returns, which says that the most benefit would come from a small set of algorithms, and adding additional algorithms won’t bring much extra improvement.

    Then they use a machine-learning model to pick the best combination of algorithms from among the 20 remaining options.

    This model is trained with a dataset specific to the user’s optimization problem, so it learns to choose algorithms that best suit the user’s particular task. Since a company like FedEx has solved routing problems many times before, using real data gleaned from past experience should lead to better solutions than starting from scratch each time.

    The model’s iterative learning process, known as contextual bandits, a form of reinforcement learning, involves picking a potential solution, getting feedback on how good it was, and then trying again to find a better solution.

    This data-driven approach accelerated MILP solvers between 30 and 70 percent without any drop in accuracy. Moreover, the speedup was similar when they applied it to a simpler, open-source solver and a more powerful, commercial solver.

    In the future, Wu and her collaborators want to apply this approach to even more complex MILP problems, where gathering labeled data to train the model could be especially challenging. Perhaps they can train the model on a smaller dataset and then tweak it to tackle a much larger optimization problem, she says. The researchers are also interested in interpreting the learned model to better understand the effectiveness of different separator algorithms.

    This research is supported, in part, by Mathworks, the National Science Foundation (NSF), the MIT Amazon Science Hub, and MIT’s Research Support Committee. More

  • in

    Technique enables AI on edge devices to keep learning over time

    Personalized deep-learning models can enable artificial intelligence chatbots that adapt to understand a user’s accent or smart keyboards that continuously update to better predict the next word based on someone’s typing history. This customization requires constant fine-tuning of a machine-learning model with new data.

    Because smartphones and other edge devices lack the memory and computational power necessary for this fine-tuning process, user data are typically uploaded to cloud servers where the model is updated. But data transmission uses a great deal of energy, and sending sensitive user data to a cloud server poses a security risk.  

    Researchers from MIT, the MIT-IBM Watson AI Lab, and elsewhere developed a technique that enables deep-learning models to efficiently adapt to new sensor data directly on an edge device.

    Their on-device training method, called PockEngine, determines which parts of a huge machine-learning model need to be updated to improve accuracy, and only stores and computes with those specific pieces. It performs the bulk of these computations while the model is being prepared, before runtime, which minimizes computational overhead and boosts the speed of the fine-tuning process.    

    When compared to other methods, PockEngine significantly sped up on-device training, performing up to 15 times faster on some hardware platforms. Moreover, PockEngine didn’t cause models to have any dip in accuracy. The researchers also found that their fine-tuning method enabled a popular AI chatbot to answer complex questions more accurately.

    “On-device fine-tuning can enable better privacy, lower costs, customization ability, and also lifelong learning, but it is not easy. Everything has to happen with a limited number of resources. We want to be able to run not only inference but also training on an edge device. With PockEngine, now we can,” says Song Han, an associate professor in the Department of Electrical Engineering and Computer Science (EECS), a member of the MIT-IBM Watson AI Lab, a distinguished scientist at NVIDIA, and senior author of an open-access paper describing PockEngine.

    Han is joined on the paper by lead author Ligeng Zhu, an EECS graduate student, as well as others at MIT, the MIT-IBM Watson AI Lab, and the University of California San Diego. The paper was recently presented at the IEEE/ACM International Symposium on Microarchitecture.

    Layer by layer

    Deep-learning models are based on neural networks, which comprise many interconnected layers of nodes, or “neurons,” that process data to make a prediction. When the model is run, a process called inference, a data input (such as an image) is passed from layer to layer until the prediction (perhaps the image label) is output at the end. During inference, each layer no longer needs to be stored after it processes the input.

    But during training and fine-tuning, the model undergoes a process known as backpropagation. In backpropagation, the output is compared to the correct answer, and then the model is run in reverse. Each layer is updated as the model’s output gets closer to the correct answer.

    Because each layer may need to be updated, the entire model and intermediate results must be stored, making fine-tuning more memory demanding than inference

    However, not all layers in the neural network are important for improving accuracy. And even for layers that are important, the entire layer may not need to be updated. Those layers, and pieces of layers, don’t need to be stored. Furthermore, one may not need to go all the way back to the first layer to improve accuracy — the process could be stopped somewhere in the middle.

    PockEngine takes advantage of these factors to speed up the fine-tuning process and cut down on the amount of computation and memory required.

    The system first fine-tunes each layer, one at a time, on a certain task and measures the accuracy improvement after each individual layer. In this way, PockEngine identifies the contribution of each layer, as well as trade-offs between accuracy and fine-tuning cost, and automatically determines the percentage of each layer that needs to be fine-tuned.

    “This method matches the accuracy very well compared to full back propagation on different tasks and different neural networks,” Han adds.

    A pared-down model

    Conventionally, the backpropagation graph is generated during runtime, which involves a great deal of computation. Instead, PockEngine does this during compile time, while the model is being prepared for deployment.

    PockEngine deletes bits of code to remove unnecessary layers or pieces of layers, creating a pared-down graph of the model to be used during runtime. It then performs other optimizations on this graph to further improve efficiency.

    Since all this only needs to be done once, it saves on computational overhead for runtime.

    “It is like before setting out on a hiking trip. At home, you would do careful planning — which trails are you going to go on, which trails are you going to ignore. So then at execution time, when you are actually hiking, you already have a very careful plan to follow,” Han explains.

    When they applied PockEngine to deep-learning models on different edge devices, including Apple M1 Chips and the digital signal processors common in many smartphones and Raspberry Pi computers, it performed on-device training up to 15 times faster, without any drop in accuracy. PockEngine also significantly slashed the amount of memory required for fine-tuning.

    The team also applied the technique to the large language model Llama-V2. With large language models, the fine-tuning process involves providing many examples, and it’s crucial for the model to learn how to interact with users, Han says. The process is also important for models tasked with solving complex problems or reasoning about solutions.

    For instance, Llama-V2 models that were fine-tuned using PockEngine answered the question “What was Michael Jackson’s last album?” correctly, while models that weren’t fine-tuned failed. PockEngine cut the time it took for each iteration of the fine-tuning process from about seven seconds to less than one second on a NVIDIA Jetson Orin, an edge GPU platform.

    In the future, the researchers want to use PockEngine to fine-tune even larger models designed to process text and images together.

    “This work addresses growing efficiency challenges posed by the adoption of large AI models such as LLMs across diverse applications in many different industries. It not only holds promise for edge applications that incorporate larger models, but also for lowering the cost of maintaining and updating large AI models in the cloud,” says Ehry MacRostie, a senior manager in Amazon’s Artificial General Intelligence division who was not involved in this study but works with MIT on related AI research through the MIT-Amazon Science Hub.

    This work was supported, in part, by the MIT-IBM Watson AI Lab, the MIT AI Hardware Program, the MIT-Amazon Science Hub, the National Science Foundation (NSF), and the Qualcomm Innovation Fellowship. More

  • in

    A more effective experimental design for engineering a cell into a new state

    A strategy for cellular reprogramming involves using targeted genetic interventions to engineer a cell into a new state. The technique holds great promise in immunotherapy, for instance, where researchers could reprogram a patient’s T-cells so they are more potent cancer killers. Someday, the approach could also help identify life-saving cancer treatments or regenerative therapies that repair disease-ravaged organs.

    But the human body has about 20,000 genes, and a genetic perturbation could be on a combination of genes or on any of the over 1,000 transcription factors that regulate the genes. Because the search space is vast and genetic experiments are costly, scientists often struggle to find the ideal perturbation for their particular application.   

    Researchers from MIT and Harvard University developed a new, computational approach that can efficiently identify optimal genetic perturbations based on a much smaller number of experiments than traditional methods.

    Their algorithmic technique leverages the cause-and-effect relationship between factors in a complex system, such as genome regulation, to prioritize the best intervention in each round of sequential experiments.

    The researchers conducted a rigorous theoretical analysis to determine that their technique did, indeed, identify optimal interventions. With that theoretical framework in place, they applied the algorithms to real biological data designed to mimic a cellular reprogramming experiment. Their algorithms were the most efficient and effective.

    “Too often, large-scale experiments are designed empirically. A careful causal framework for sequential experimentation may allow identifying optimal interventions with fewer trials, thereby reducing experimental costs,” says co-senior author Caroline Uhler, a professor in the Department of Electrical Engineering and Computer Science (EECS) 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 Institute for Data, Systems and Society (IDSS).

    Joining Uhler on the paper, which appears today in Nature Machine Intelligence, are lead author Jiaqi Zhang, a graduate student and Eric and Wendy Schmidt Center Fellow; co-senior author Themistoklis P. Sapsis, professor of mechanical and ocean engineering at MIT and a member of IDSS; and others at Harvard and MIT.

    Active learning

    When scientists try to design an effective intervention for a complex system, like in cellular reprogramming, they often perform experiments sequentially. Such settings are ideally suited for the use of a machine-learning approach called active learning. Data samples are collected and used to learn a model of the system that incorporates the knowledge gathered so far. From this model, an acquisition function is designed — an equation that evaluates all potential interventions and picks the best one to test in the next trial.

    This process is repeated until an optimal intervention is identified (or resources to fund subsequent experiments run out).

    “While there are several generic acquisition functions to sequentially design experiments, these are not effective for problems of such complexity, leading to very slow convergence,” Sapsis explains.

    Acquisition functions typically consider correlation between factors, such as which genes are co-expressed. But focusing only on correlation ignores the regulatory relationships or causal structure of the system. For instance, a genetic intervention can only affect the expression of downstream genes, but a correlation-based approach would not be able to distinguish between genes that are upstream or downstream.

    “You can learn some of this causal knowledge from the data and use that to design an intervention more efficiently,” Zhang explains.

    The MIT and Harvard researchers leveraged this underlying causal structure for their technique. First, they carefully constructed an algorithm so it can only learn models of the system that account for causal relationships.

    Then the researchers designed the acquisition function so it automatically evaluates interventions using information on these causal relationships. They crafted this function so it prioritizes the most informative interventions, meaning those most likely to lead to the optimal intervention in subsequent experiments.

    “By considering causal models instead of correlation-based models, we can already rule out certain interventions. Then, whenever you get new data, you can learn a more accurate causal model and thereby further shrink the space of interventions,” Uhler explains.

    This smaller search space, coupled with the acquisition function’s special focus on the most informative interventions, is what makes their approach so efficient.

    The researchers further improved their acquisition function using a technique known as output weighting, inspired by the study of extreme events in complex systems. This method carefully emphasizes interventions that are likely to be closer to the optimal intervention.

    “Essentially, we view an optimal intervention as an ‘extreme event’ within the space of all possible, suboptimal interventions and use some of the ideas we have developed for these problems,” Sapsis says.    

    Enhanced efficiency

    They tested their algorithms using real biological data in a simulated cellular reprogramming experiment. For this test, they sought a genetic perturbation that would result in a desired shift in average gene expression. Their acquisition functions consistently identified better interventions than baseline methods through every step in the multi-stage experiment.

    “If you cut the experiment off at any stage, ours would still be more efficient than the baselines. This means you could run fewer experiments and get the same or better results,” Zhang says.

    The researchers are currently working with experimentalists to apply their technique toward cellular reprogramming in the lab.

    Their approach could also be applied to problems outside genomics, such as identifying optimal prices for consumer products or enabling optimal feedback control in fluid mechanics applications.

    In the future, they plan to enhance their technique for optimizations beyond those that seek to match a desired mean. In addition, their method assumes that scientists already understand the causal relationships in their system, but future work could explore how to use AI to learn that information, as well.

    This work was funded, in part, by the Office of Naval Research, the MIT-IBM Watson AI Lab, the MIT J-Clinic for Machine Learning and Health, the Eric and Wendy Schmidt Center at the Broad Institute, a Simons Investigator Award, the Air Force Office of Scientific Research, and a National Science Foundation Graduate Fellowship. More

  • in

    Helping computer vision and language models understand what they see

    Powerful machine-learning algorithms known as vision and language models, which learn to match text with images, have shown remarkable results when asked to generate captions or summarize videos.

    While these models excel at identifying objects, they often struggle to understand concepts, like object attributes or the arrangement of items in a scene. For instance, a vision and language model might recognize the cup and table in an image, but fail to grasp that the cup is sitting on the table.

    Researchers from MIT, the MIT-IBM Watson AI Lab, and elsewhere have demonstrated a new technique that utilizes computer-generated data to help vision and language models overcome this shortcoming.

    The researchers created a synthetic dataset of images that depict a wide range of scenarios, object arrangements, and human actions, coupled with detailed text descriptions. They used this annotated dataset to “fix” vision and language models so they can learn concepts more effectively. Their technique ensures these models can still make accurate predictions when they see real images.

    When they tested models on concept understanding, the researchers found that their technique boosted accuracy by up to 10 percent. This could improve systems that automatically caption videos or enhance models that provide natural language answers to questions about images, with applications in fields like e-commerce or health care.

    “With this work, we are going beyond nouns in the sense that we are going beyond just the names of objects to more of the semantic concept of an object and everything around it. Our idea was that, when a machine-learning model sees objects in many different arrangements, it will have a better idea of how arrangement matters in a scene,” says Khaled Shehada, a graduate student in the Department of Electrical Engineering and Computer Science and co-author of a paper on this technique.

    Shehada wrote the paper with lead author Paola Cascante-Bonilla, a computer science graduate student at Rice University; Aude Oliva, director of strategic industry engagement at the MIT Schwarzman College of Computing, MIT director of the MIT-IBM Watson AI Lab, and a senior research scientist in the Computer Science and Artificial Intelligence Laboratory (CSAIL); senior author Leonid Karlinsky, a research staff member in the MIT-IBM Watson AI Lab; and others at MIT, the MIT-IBM Watson AI Lab, Georgia Tech, Rice University, École des Ponts, Weizmann Institute of Science, and IBM Research. The paper will be presented at the International Conference on Computer Vision.

    Focusing on objects

    Vision and language models typically learn to identify objects in a scene, and can end up ignoring object attributes, such as color and size, or positional relationships, such as which object is on top of another object.

    This is due to the method with which these models are often trained, known as contrastive learning. This training method involves forcing a model to predict the correspondence between images and text. When comparing natural images, the objects in each scene tend to cause the most striking differences. (Perhaps one image shows a horse in a field while the second shows a sailboat on the water.)

    “Every image could be uniquely defined by the objects in the image. So, when you do contrastive learning, just focusing on the nouns and objects would solve the problem. Why would the model do anything differently?” says Karlinsky.

    The researchers sought to mitigate this problem by using synthetic data to fine-tune a vision and language model. The fine-tuning process involves tweaking a model that has already been trained to improve its performance on a specific task.

    They used a computer to automatically create synthetic videos with diverse 3D environments and objects, such as furniture and luggage, and added human avatars that interacted with the objects.

    Using individual frames of these videos, they generated nearly 800,000 photorealistic images, and then paired each with a detailed caption. The researchers developed a methodology for annotating every aspect of the image to capture object attributes, positional relationships, and human-object interactions clearly and consistently in dense captions.

    Because the researchers created the images, they could control the appearance and position of objects, as well as the gender, clothing, poses, and actions of the human avatars.

    “Synthetic data allows a lot of diversity. With real images, you might not have a lot of elephants in a room, but with synthetic data, you could actually have a pink elephant in a room with a human, if you want,” Cascante-Bonilla says.

    Synthetic data have other advantages, too. They are cheaper to generate than real data, yet the images are highly photorealistic. They also preserve privacy because no real humans are shown in the images. And, because data are produced automatically by a computer, they can be generated quickly in massive quantities.

    By using different camera viewpoints, or slightly changing the positions or attributes of objects, the researchers created a dataset with a far wider variety of scenarios than one would find in a natural dataset.

    Fine-tune, but don’t forget

    However, when one fine-tunes a model with synthetic data, there is a risk that model might “forget” what it learned when it was originally trained with real data.

    The researchers employed a few techniques to prevent this problem, such as adjusting the synthetic data so colors, lighting, and shadows more closely match those found in natural images. They also made adjustments to the model’s inner-workings after fine-tuning to further reduce any forgetfulness.

    Their synthetic dataset and fine-tuning strategy improved the ability of popular vision and language models to accurately recognize concepts by up to 10 percent. At the same time, the models did not forget what they had already learned.

    Now that they have shown how synthetic data can be used to solve this problem, the researchers want to identify ways to improve the visual quality and diversity of these data, as well as the underlying physics that makes synthetic scenes look realistic. In addition, they plan to test the limits of scalability, and investigate whether model improvement starts to plateau with larger and more diverse synthetic datasets.

    This research is funded, in part, by the U.S. Defense Advanced Research Projects Agency, the National Science Foundation, and the MIT-IBM Watson AI Lab. More