More stories

  • in

    Computers that power self-driving cars could be a huge driver of global carbon emissions

    In the future, the energy needed to run the powerful computers on board a global fleet of autonomous vehicles could generate as many greenhouse gas emissions as all the data centers in the world today.

    That is one key finding of a new study from MIT researchers that explored the potential energy consumption and related carbon emissions if autonomous vehicles are widely adopted.

    The data centers that house the physical computing infrastructure used for running applications are widely known for their large carbon footprint: They currently account for about 0.3 percent of global greenhouse gas emissions, or about as much carbon as the country of Argentina produces annually, according to the International Energy Agency. Realizing that less attention has been paid to the potential footprint of autonomous vehicles, the MIT researchers built a statistical model to study the problem. They determined that 1 billion autonomous vehicles, each driving for one hour per day with a computer consuming 840 watts, would consume enough energy to generate about the same amount of emissions as data centers currently do.

    The researchers also found that in over 90 percent of modeled scenarios, to keep autonomous vehicle emissions from zooming past current data center emissions, each vehicle must use less than 1.2 kilowatts of power for computing, which would require more efficient hardware. In one scenario — where 95 percent of the global fleet of vehicles is autonomous in 2050, computational workloads double every three years, and the world continues to decarbonize at the current rate — they found that hardware efficiency would need to double faster than every 1.1 years to keep emissions under those levels.

    “If we just keep the business-as-usual trends in decarbonization and the current rate of hardware efficiency improvements, it doesn’t seem like it is going to be enough to constrain the emissions from computing onboard autonomous vehicles. This has the potential to become an enormous problem. But if we get ahead of it, we could design more efficient autonomous vehicles that have a smaller carbon footprint from the start,” says first author Soumya Sudhakar, a graduate student in aeronautics and astronautics.

    Sudhakar wrote the paper with her co-advisors Vivienne Sze, associate professor in the Department of Electrical Engineering and Computer Science (EECS) and a member of the Research Laboratory of Electronics (RLE); and Sertac Karaman, associate professor of aeronautics and astronautics and director of the Laboratory for Information and Decision Systems (LIDS). The research appears today in the January-February issue of IEEE Micro.

    Modeling emissions

    The researchers built a framework to explore the operational emissions from computers on board a global fleet of electric vehicles that are fully autonomous, meaning they don’t require a back-up human driver.

    The model is a function of the number of vehicles in the global fleet, the power of each computer on each vehicle, the hours driven by each vehicle, and the carbon intensity of the electricity powering each computer.

    “On its own, that looks like a deceptively simple equation. But each of those variables contains a lot of uncertainty because we are considering an emerging application that is not here yet,” Sudhakar says.

    For instance, some research suggests that the amount of time driven in autonomous vehicles might increase because people can multitask while driving and the young and the elderly could drive more. But other research suggests that time spent driving might decrease because algorithms could find optimal routes that get people to their destinations faster.

    In addition to considering these uncertainties, the researchers also needed to model advanced computing hardware and software that doesn’t exist yet.

    To accomplish that, they modeled the workload of a popular algorithm for autonomous vehicles, known as a multitask deep neural network because it can perform many tasks at once. They explored how much energy this deep neural network would consume if it were processing many high-resolution inputs from many cameras with high frame rates, simultaneously.

    When they used the probabilistic model to explore different scenarios, Sudhakar was surprised by how quickly the algorithms’ workload added up.

    For example, if an autonomous vehicle has 10 deep neural networks processing images from 10 cameras, and that vehicle drives for one hour a day, it will make 21.6 million inferences each day. One billion vehicles would make 21.6 quadrillion inferences. To put that into perspective, all of Facebook’s data centers worldwide make a few trillion inferences each day (1 quadrillion is 1,000 trillion).

    “After seeing the results, this makes a lot of sense, but it is not something that is on a lot of people’s radar. These vehicles could actually be using a ton of computer power. They have a 360-degree view of the world, so while we have two eyes, they may have 20 eyes, looking all over the place and trying to understand all the things that are happening at the same time,” Karaman says.

    Autonomous vehicles would be used for moving goods, as well as people, so there could be a massive amount of computing power distributed along global supply chains, he says. And their model only considers computing — it doesn’t take into account the energy consumed by vehicle sensors or the emissions generated during manufacturing.

    Keeping emissions in check

    To keep emissions from spiraling out of control, the researchers found that each autonomous vehicle needs to consume less than 1.2 kilowatts of energy for computing. For that to be possible, computing hardware must become more efficient at a significantly faster pace, doubling in efficiency about every 1.1 years.

    One way to boost that efficiency could be to use more specialized hardware, which is designed to run specific driving algorithms. Because researchers know the navigation and perception tasks required for autonomous driving, it could be easier to design specialized hardware for those tasks, Sudhakar says. But vehicles tend to have 10- or 20-year lifespans, so one challenge in developing specialized hardware would be to “future-proof” it so it can run new algorithms.

    In the future, researchers could also make the algorithms more efficient, so they would need less computing power. However, this is also challenging because trading off some accuracy for more efficiency could hamper vehicle safety.

    Now that they have demonstrated this framework, the researchers want to continue exploring hardware efficiency and algorithm improvements. In addition, they say their model can be enhanced by characterizing embodied carbon from autonomous vehicles — the carbon emissions generated when a car is manufactured — and emissions from a vehicle’s sensors.

    While there are still many scenarios to explore, the researchers hope that this work sheds light on a potential problem people may not have considered.

    “We are hoping that people will think of emissions and carbon efficiency as important metrics to consider in their designs. The energy consumption of an autonomous vehicle is really critical, not just for extending the battery life, but also for sustainability,” says Sze.

    This research was funded, in part, by the National Science Foundation and the MIT-Accenture Fellowship. More

  • in

    New MIT internships expand research opportunities in Africa

    With new support from the Office of the Associate Provost for International Activities, MIT International Science and Technology Initiatives (MISTI) and the MIT-Africa program are expanding internship opportunities for MIT students at universities and leading academic research centers in Africa. This past summer, MISTI supported 10 MIT student interns at African universities, significantly more than in any previous year.

    “These internships are an opportunity to better merge the research ecosystem of MIT with academia-based research systems in Africa,” says Evan Lieberman, the Total Professor of Political Science and Contemporary Africa and faculty director for MISTI.

    For decades, MISTI has helped MIT students to learn and explore through international experiential learning opportunities and internships in industries like health care, education, agriculture, and energy. MISTI’s MIT-Africa Seed Fund supports collaborative research between MIT faculty and Africa-based researchers, and the new student research internship opportunities are part of a broader vision for deeper engagement between MIT and research institutions across the African continent.

    While Africa is home to 12.5 percent of the world’s population, it generates less than 1 percent of scientific research output in the form of academic journal publications, according to the African Academy of Sciences. Research internships are one way that MIT can build mutually beneficial partnerships across Africa’s research ecosystem, to advance knowledge and spawn innovation in fields important to MIT and its African counterparts, including health care, biotechnology, urban planning, sustainable energy, and education.

    Ari Jacobovits, managing director of MIT-Africa, notes that the new internships provide additional funding to the lab hosting the MIT intern, enabling them to hire a counterpart student research intern from the local university. This support can make the internships more financially feasible for host institutions and helps to grow the research pipeline.

    With the support of MIT, State University of Zanzibar (SUZA) lecturers Raya Ahmada and Abubakar Bakar were able to hire local students to work alongside MIT graduate students Mel Isidor and Rajan Hoyle. Together the students collaborated over a summer on a mapping project designed to plan and protect Zanzibar’s coastal economy.

    “It’s been really exciting to work with research peers in a setting where we can all learn alongside one another and develop this project together,” says Hoyle.

    Using low-cost drone technology, the students and their local counterparts worked to create detailed maps of Zanzibar to support community planning around resilience projects designed to combat coastal flooding and deforestation and assess climate-related impacts to seaweed farming activities. 

    “I really appreciated learning about how engagement happens in this particular context and how community members understand local environmental challenges and conditions based on research and lived experience,” says Isidor. “This is beneficial for us whether we’re working in an international context or in the United States.”

    For biology major Shaida Nishat, her internship at the University of Cape Town allowed her to work in a vital sphere of public health and provided her with the chance to work with a diverse, international team headed by Associate Professor Salome Maswine, head of the global surgery division and a widely-renowned expert in global surgery, a multidisciplinary field in the sphere of global health focused on improved and equitable surgical outcomes.

    “It broadened my perspective as to how an effort like global surgery ties so many nations together through a common goal that would benefit them all,” says Nishat, who plans to pursue a career in public health.

    For computer science sophomore Antonio L. Ortiz Bigio, the MISTI research internship in Africa was an incomparable experience, culturally and professionally. Bigio interned at the Robotics Autonomous Intelligence and Learning Laboratory at the University of Witwatersrand in Johannesburg, led by Professor Benjamin Rosman, where he developed software to enable a robot to play chess. The experience has inspired Bigio to continue to pursue robotics and machine learning.

    Participating faculty at the host institutions welcomed their MIT interns, and were impressed by their capabilities. Both Rosman and Maswime described their MIT interns as hard-working and valued team members, who had helped to advance their own work.  

    Building strong global partnerships, whether through faculty research, student internships, or other initiatives, takes time and cultivation, explains Jacobovits. Each successful collaboration helps to seed future exchanges and builds interest at MIT and peer institutions in creative partnerships. As MIT continues to deepen its connections to institutions and researchers across Africa, says Jacobovits, “students like Shaida, Rajan, Mel, and Antonio are really effective ambassadors in building those networks.” More

  • in

    Ian Hutchinson: A lifetime probing plasma, on Earth and in space

    Ordinary folks gazing at the night sky can readily spot Earth’s close neighbors and the light of distant stars. But when Ian Hutchinson scans the cosmos, he takes in a great deal more. There is, for instance, the constant rush of plasma — highly charged ionized gases — from the sun. As this plasma flows by solid bodies such as the moon, it interacts with them electromagnetically, sometimes generating a phenomenon called an electron hole — a perturbation in the gaseous solar tide that forms a solitary, long-lived wave. Hutchinson, a professor in the MIT Department of Nuclear Science and Engineering (NSE), knows they exist because he found a way to measure them.

    “When I look up at the moon with my sweetheart, my wife of 48 years, I imagine that streaming from its dark side are electron holes that my students and I predicted, and that we then discovered,” he says. “It’s quite sentimental to me.”

    Hutchinson’s studies of these wave phenomena, summed up in a paper, “Electron holes in phase space: What they are and why they matter,” recently earned the 2022 Ronald C. Davidson Award for Plasma Physics presented by the American Physical Society’s Division of Plasma Physics.

    Measuring perturbations in plasma

    Hutchinson’s exploration of electron holes was sparked by his work over many decades in fusion energy, another branch of plasma physics. He has made many contributions to the design, operation, and experimental investigation of tokamaks — a toroidal magnetic confinement device — intended to replicate and harness the fiery thermonuclear reactions in the plasma of stars for carbon-free energy on Earth. Hutchinson took a particular interest in how to measure the plasma, notably the flow at the edges of tokamaks.

    Heat generated from fusion reactions may escape magnetic confinement and build up along these edges, leading to potential temperature spikes that impact the performance of the confinement device. Hutchinson discovered how to interpret signals from small probes to measure and track plasma velocity at the tokamak’s edge.

    “My theoretical work also showed that these probes quite likely induce electron holes,” he says. But proving this contention required experiments at resolutions in time and space beyond what tokamaks allow. That’s when Hutchinson had an important insight.

    “I realized that the phenomena we were trying to investigate can actually be measured with exquisite accuracy by satellites that travel through plasma surrounding Earth and other solid bodies,” he says. Although plasmas in space are at a much larger scale than the plasmas generated in the laboratory, measurements of these gases by a satellite is analogous “to a situation where we fly a tiny micron-sized spacecraft through the wakes of probes at the edge of tokamaks,” says Hutchinson.

    Using satellite data provided by NASA, Hutchinson set about analyzing solar plasma as it whips by the moon. “We predicted instabilities and the generation of electron holes,” he recounts. “Our theory passed with flying colors: We saw lots of holes in the wake of the moon, and few elsewhere.”

    Developing tokamaks

    Hutchinson grew up in the English midlands and attended Cambridge University, where he became “intrigued by plasma physics in a course taught by an entertaining and effective teacher,” he says.

    Hutchinson headed for doctoral studies at Australian National University on fellowship. The experience afforded him his first opportunity for research on plasma confinement. “There I was at the ends of the Earth, and I was one of very few scientists worldwide with a tokamak almost to myself,” he says. “It was a device that had risen to the top of everyone’s agenda in fusion research as something we really needed to understand.”

    His dissertation, which examined instabilities in plasma, and his hands-on experience with the device, brought him to the attention of Ronald Parker SM ’63, PhD ’67, now emeritus professor of nuclear science and engineering and electrical engineering and computer science, who was building MIT’s Alcator tokamak program.

    In 1976, Hutchinson joined this group, spending three years as a research scientist. After an interval in Britain, he returned to MIT with a faculty position in NSE, and soon, a leadership role in developing the next phase of the Institute’s fusion experiment, the Alcator-C Mod tokamak.

    “This was a major development of the high-magnetic field approach to fusion,” says Hutchinson. Powerful magnets are essential for containing the superhot plasma; the MIT group developed an experiment with a magnetic field more than 150,000 times the strength of the Earth’s magnetic field. “We were in the business of determining whether tokamaks had sufficiently good confinement to function as fusion reactors,” he says.

    Hutchinson oversaw the nearly six-year construction of the device, which was funded by the U.S. Department of Energy. He then led its operation starting in 1993, creating a national facility for experiments that drew scientists and students from around the world. At the time, it was the largest research group on campus at MIT.

    In their studies, scientists employed novel heating and sustainment techniques using radio waves and microwaves. They also discovered new methods for performing diagnostics inside the tokamak. “Alcator C-Mod demonstrated excellent confinement in a more compact and cost-effective device,” says Hutchinson. “It was unique in the world.”

    Hutchinson is proud of Alcator C-Mod’s technological achievements, including its record for highest plasma pressure for a magnetic confinement device. But this large-scale project holds even greater significance for him. “Alcator C-Mod helped beat a new path in fusion research, and has become the basis for the SPARC tokamak now under construction,” he says.

    SPARC is a compact, high-magnetic field fusion energy device under development through a collaboration between MIT’s Plasma Science and Fusion Center and startup Commonwealth Fusions Systems. Its goal is to demonstrate net energy gain from fusion, prove the viability of fusion as a source of carbon-free energy, and tip the scales in the race against climate change. A number of SPARC’s leaders are students Hutchinson taught. “This is a source of considerable satisfaction,” he says. “Some of their down-to-Earth realism comes from me, and perhaps some of their aspirations have been molded by their work with me.” 

    A new phase

    After leading Alcator C-Mod for 15 years and generating hundreds of journal articles, Hutchinson served as NSE’s department head from 2003 to 2009. He wrote the standard textbook on measuring plasmas, and has more recently written “A Student’s Guide to Numerical Methods” (2015), which evolved from a course he taught to introduce graduate students to computational problem-solving in physics and engineering.

    After this, his 40th year on the MIT faculty, Hutchinson will be stepping back from teaching. “It’s important for new generations of students to be taught by people at the pinnacle of their mental and intellectual capacity, and when you reach my age, you’re aware of the fact that you’re slowing down,” he says.

    Hutchinson’s at no loss for ways to spend his time. As a devout Christian, he speaks and writes about the relationship between religion and science, trying to help skeptics on both sides find common ground. He sings in two choral groups, and is very busy grandparenting four grandsons. For a complete change of pace, Hutchinson goes fly fishing.

    But he still has plans to explore new frontiers in plasma physics. “I’m gratified to say I still do important research,” he says. “I’ve solved most of the problems in electron holes, and now I need to say something about ion holes!” More

  • in

    Machine learning facilitates “turbulence tracking” in fusion reactors

    Fusion, which promises practically unlimited, carbon-free energy using the same processes that power the sun, is at the heart of a worldwide research effort that could help mitigate climate change.

    A multidisciplinary team of researchers is now bringing tools and insights from machine learning to aid this effort. Scientists from MIT and elsewhere have used computer-vision models to identify and track turbulent structures that appear under the conditions needed to facilitate fusion reactions.

    Monitoring the formation and movements of these structures, called filaments or “blobs,” is important for understanding the heat and particle flows exiting from the reacting fuel, which ultimately determines the engineering requirements for the reactor walls to meet those flows. However, scientists typically study blobs using averaging techniques, which trade details of individual structures in favor of aggregate statistics. Individual blob information must be tracked by marking them manually in video data. 

    The researchers built a synthetic video dataset of plasma turbulence to make this process more effective and efficient. They used it to train four computer vision models, each of which identifies and tracks blobs. They trained the models to pinpoint blobs in the same ways that humans would.

    When the researchers tested the trained models using real video clips, the models could identify blobs with high accuracy — more than 80 percent in some cases. The models were also able to effectively estimate the size of blobs and the speeds at which they moved.

    Because millions of video frames are captured during just one fusion experiment, using machine-learning models to track blobs could give scientists much more detailed information.

    “Before, we could get a macroscopic picture of what these structures are doing on average. Now, we have a microscope and the computational power to analyze one event at a time. If we take a step back, what this reveals is the power available from these machine-learning techniques, and ways to use these computational resources to make progress,” says Theodore Golfinopoulos, a research scientist at the MIT Plasma Science and Fusion Center and co-author of a paper detailing these approaches.

    His fellow co-authors include lead author Woonghee “Harry” Han, a physics PhD candidate; senior author Iddo Drori, a visiting professor in the Computer Science and Artificial Intelligence Laboratory (CSAIL), faculty associate professor at Boston University, and adjunct at Columbia University; as well as others from the MIT Plasma Science and Fusion Center, the MIT Department of Civil and Environmental Engineering, and the Swiss Federal Institute of Technology at Lausanne in Switzerland. The research appears today in Nature Scientific Reports.

    Heating things up

    For more than 70 years, scientists have sought to use controlled thermonuclear fusion reactions to develop an energy source. To reach the conditions necessary for a fusion reaction, fuel must be heated to temperatures above 100 million degrees Celsius. (The core of the sun is about 15 million degrees Celsius.)

    A common method for containing this super-hot fuel, called plasma, is to use a tokamak. These devices utilize extremely powerful magnetic fields to hold the plasma in place and control the interaction between the exhaust heat from the plasma and the reactor walls.

    However, blobs appear like filaments falling out of the plasma at the very edge, between the plasma and the reactor walls. These random, turbulent structures affect how energy flows between the plasma and the reactor.

    “Knowing what the blobs are doing strongly constrains the engineering performance that your tokamak power plant needs at the edge,” adds Golfinopoulos.

    Researchers use a unique imaging technique to capture video of the plasma’s turbulent edge during experiments. An experimental campaign may last months; a typical day will produce about 30 seconds of data, corresponding to roughly 60 million video frames, with thousands of blobs appearing each second. This makes it impossible to track all blobs manually, so researchers rely on average sampling techniques that only provide broad characteristics of blob size, speed, and frequency.

    “On the other hand, machine learning provides a solution to this by blob-by-blob tracking for every frame, not just average quantities. This gives us much more knowledge about what is happening at the boundary of the plasma,” Han says.

    He and his co-authors took four well-established computer vision models, which are commonly used for applications like autonomous driving, and trained them to tackle this problem.

    Simulating blobs

    To train these models, they created a vast dataset of synthetic video clips that captured the blobs’ random and unpredictable nature.

    “Sometimes they change direction or speed, sometimes multiple blobs merge, or they split apart. These kinds of events were not considered before with traditional approaches, but we could freely simulate those behaviors in the synthetic data,” Han says.

    Creating synthetic data also allowed them to label each blob, which made the training process more effective, Drori adds.

    Using these synthetic data, they trained the models to draw boundaries around blobs, teaching them to closely mimic what a human scientist would draw.

    Then they tested the models using real video data from experiments. First, they measured how closely the boundaries the models drew matched up with actual blob contours.

    But they also wanted to see if the models predicted objects that humans would identify. They asked three human experts to pinpoint the centers of blobs in video frames and checked to see if the models predicted blobs in those same locations.

    The models were able to draw accurate blob boundaries, overlapping with brightness contours which are considered ground-truth, about 80 percent of the time. Their evaluations were similar to those of human experts, and successfully predicted the theory-defined regime of the blob, which agrees with the results from a traditional method.

    Now that they have shown the success of using synthetic data and computer vision models for tracking blobs, the researchers plan to apply these techniques to other problems in fusion research, such as estimating particle transport at the boundary of a plasma, Han says.

    They also made the dataset and models publicly available, and look forward to seeing how other research groups apply these tools to study the dynamics of blobs, says Drori.

    “Prior to this, there was a barrier to entry that mostly the only people working on this problem were plasma physicists, who had the datasets and were using their methods. There is a huge machine-learning and computer-vision community. One goal of this work is to encourage participation in fusion research from the broader machine-learning community toward the broader goal of helping solve the critical problem of climate change,” he adds.

    This research is supported, in part, by the U.S. Department of Energy and the Swiss National Science Foundation. More

  • in

    Scientists chart how exercise affects the body

    Exercise is well-known to help people lose weight and avoid gaining it. However, identifying the cellular mechanisms that underlie this process has proven difficult because so many cells and tissues are involved.

    In a new study in mice that expands researchers’ understanding of how exercise and diet affect the body, MIT and Harvard Medical School researchers have mapped out many of the cells, genes, and cellular pathways that are modified by exercise or high-fat diet. The findings could offer potential targets for drugs that could help to enhance or mimic the benefits of exercise, the researchers say.

    “It is extremely important to understand the molecular mechanisms that are drivers of the beneficial effects of exercise and the detrimental effects of a high-fat diet, so that we can understand how we can intervene, and develop drugs that mimic the impact of exercise across multiple tissues,” says Manolis Kellis, a professor of computer science in MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and a member of the Broad Institute of MIT and Harvard.

    The researchers studied mice with high-fat or normal diets, who were either sedentary or given the opportunity to exercise whenever they wanted. Using single-cell RNA sequencing, the researchers cataloged the responses of 53 types of cells found in skeletal muscle and two types of fatty tissue.

    “One of the general points that we found in our study, which is overwhelmingly clear, is how high-fat diets push all of these cells and systems in one way, and exercise seems to be pushing them nearly all in the opposite way,” Kellis says. “It says that exercise can really have a major effect throughout the body.”

    Kellis and Laurie Goodyear, a professor of medicine at Harvard Medical School and senior investigator at the Joslin Diabetes Center, are the senior authors of the study, which appears today in the journal Cell Metabolism. Jiekun Yang, a research scientist in MIT CSAIL; Maria Vamvini, an instructor of medicine at the Joslin Diabetes Center; and Pasquale Nigro, an instructor of medicine at the Joslin Diabetes Center, are the lead authors of the paper.

    The risks of obesity

    Obesity is a growing health problem around the world. In the United States, more than 40 percent of the population is considered obese, and nearly 75 percent is overweight. Being overweight is a risk factor for many diseases, including heart disease, cancer, Alzheimer’s disease, and even infectious diseases such as Covid-19.

    “Obesity, along with aging, is a global factor that contributes to every aspect of human health,” Kellis says.

    Several years ago, his lab performed a study on the FTO gene region, which has been strongly linked to obesity risk. In that 2015 study, the research team found that genes in this region control a pathway that prompts immature fat cells called progenitor adipocytes to either become fat-burning cells or fat-storing cells.

    That finding, which demonstrated a clear genetic component to obesity, motivated Kellis to begin looking at how exercise, a well-known behavioral intervention that can prevent obesity, might act on progenitor adipocytes at the cellular level.

    To explore that question, Kellis and his colleagues decided to perform single-cell RNA sequencing of three types of tissue — skeletal muscle, visceral white adipose tissue (found packed around internal organs, where it stores fat), and subcutaneous white adipose tissue (which is found under the skin and primarily burns fat).

    These tissues came from mice from four different experimental groups. For three weeks, two groups of mice were fed either a normal diet or a high-fat diet. For the next three weeks, each of those two groups were further divided into a sedentary group and an exercise group, which had continuous access to a treadmill.

    By analyzing tissues from those mice, the researchers were able to comprehensively catalog the genes that were activated or suppressed by exercise in 53 different cell types.

    The researchers found that in all three tissue types, mesenchymal stem cells (MSCs) appeared to control many of the diet and exercise-induced effects that they observed. MSCs are stem cells that can differentiate into other cell types, including fat cells and fibroblasts. In adipose tissue, the researchers found that a high-fat diet modulated MSCs’ capacity to differentiate into fat-storing cells, while exercise reversed this effect.

    In addition to promoting fat storage, the researchers found that a high-fat diet also stimulated MSCs to secrete factors that remodel the extracellular matrix (ECM) — a network of proteins and other molecules that surround and support cells and tissues in the body. This ECM remodeling helps provide structure for enlarged fat-storing cells and also creates a more inflammatory environment.

    “As the adipocytes become overloaded with lipids, there’s an extreme amount of stress, and that causes low-grade inflammation, which is systemic and preserved for a long time,” Kellis says. “That is one of the factors that is contributing to many of the adverse effects of obesity.”

    Circadian effects

    The researchers also found that high-fat diets and exercise had opposing effects on cellular pathways that control circadian rhythms — the 24-hour cycles that govern many functions, from sleep to body temperature, hormone release, and digestion. The study revealed that exercise boosts the expression of genes that regulate these rhythms, while a high-fat diet suppresses them.

    “There have been a lot of studies showing that when you eat during the day is extremely important in how you absorb the calories,” Kellis says. “The circadian rhythm connection is a very important one, and shows how obesity and exercise are in fact directly impacting that circadian rhythm in peripheral organs, which could act systemically on distal clocks and regulate stem cell functions and immunity.”

    The researchers then compared their results to a database of human genes that have been linked with metabolic traits. They found that two of the circadian rhythm genes they identified in this study, known as DBP and CDKN1A, have genetic variants that have been associated with a higher risk of obesity in humans.

    “These results help us see the translational values of these targets, and how we could potentially target specific biological processes in specific cell types,” Yang says.

    The researchers are now analyzing samples of small intestine, liver, and brain tissue from the mice in this study, to explore the effects of exercise and high-fat diets on those tissues. They are also conducting work with human volunteers to sample blood and biopsies and study similarities and differences between human and mouse physiology. They hope that their findings will help guide drug developers in designing drugs that might mimic some of the beneficial effects of exercise.

    “The message for everyone should be, eat a healthy diet and exercise if possible,” Kellis says. “For those for whom this is not possible, due to low access to healthy foods, or due to disabilities or other factors that prevent exercise, or simply lack of time to have a healthy diet or a healthy lifestyle, what this study says is that we now have a better handle on the pathways, the specific genes, and the specific molecular and cellular processes that we should be manipulating therapeutically.”

    The research was funded by the National Institutes of Health and the Novo Nordisk Research Center in Seattle. More

  • in

    MIT students contribute to success of historic fusion experiment

    For more than half a century, researchers around the world have been engaged in attempts to achieve fusion ignition in a laboratory, a grand challenge of the 21st century. The High-Energy-Density Physics (HEDP) group at MIT’s Plasma Science and Fusion Center has focused on an approach called inertial confinement fusion (ICF), which uses lasers to implode a pellet of fuel in a quest for ignition. This group, including nine former and current MIT students, was crucial to an historic ICF ignition experiment performed in 2021; the results were published on the anniversary of that success.

    On Aug. 8, 2021, researchers at the National Ignition Facility (NIF), Lawrence Livermore National Laboratory (LLNL), used 192 laser beams to illuminate the inside of a tiny gold cylinder encapsulating a spherical capsule filled with deuterium-tritium fuel in their quest to produce significant fusion energy. Although researchers had followed this process many times before, using different parameters, this time the ensuing implosion produced an historic fusion yield of 1.37 megaJoules, as measured by a suite of neutron diagnostics. These included the MIT-developed and analyzed Magnetic Recoil Spectrometer (MRS). This result was published in Physical Review Letters on Aug. 8, the one-year anniversary of the ground-breaking development, unequivocally indicating that the first controlled fusion experiment reached ignition.

    Governed by the Lawson criterion, a plasma ignites when the internal fusion heating power is high enough to overcome the physical processes that cool the fusion plasma, creating a positive thermodynamic feedback loop that very rapidly increases the plasma temperature. In the case of ICF, ignition is a state where the fusion plasma can initiate a “fuel burn propagation” into the surrounding dense and cold fuel, enabling the possibility of high fusion-energy gain.

    “This historic result certainly demonstrates that the ignition threshold is a real concept, with well-predicted theoretical calculations, and that a fusion plasma can be ignited in a laboratory” says HEDP Division Head Johan Frenje.

    The HEDP division has contributed to the success of the ignition program at the NIF for more than a decade by providing and using a dozen diagnostics, implemented by MIT PhD students and staff, which have been critical for assessing the performance of an implosion. The hundreds of co-authors on the paper attest to the collaborative effort that went into this milestone. MIT’s contributors included the only student co-authors.

    “The students are responsible for implementing and using a diagnostic to obtain data important to the ICF program at the NIF, says Frenje. “Being responsible for running a diagnostic at the NIF has allowed them to actively participate in the scientific dialog and thus get directly exposed to cutting-edge science.”

    Students involved from the MIT Department of Physics were Neel Kabadi, Graeme Sutcliffe, Tim Johnson, Jacob Pearcy, and Ben Reichelt; students from the Department of Nuclear Science and Engineering included Brandon Lahmann, Patrick Adrian, and Justin Kunimune.

    In addition, former student Alex Zylstra PhD ’15, now a physicist at LLNL, was the experimental lead of this record implosion experiment. More

  • in

    Taking a magnifying glass to data center operations

    When the MIT Lincoln Laboratory Supercomputing Center (LLSC) unveiled its TX-GAIA supercomputer in 2019, it provided the MIT community a powerful new resource for applying artificial intelligence to their research. Anyone at MIT can submit a job to the system, which churns through trillions of operations per second to train models for diverse applications, such as spotting tumors in medical images, discovering new drugs, or modeling climate effects. But with this great power comes the great responsibility of managing and operating it in a sustainable manner — and the team is looking for ways to improve.

    “We have these powerful computational tools that let researchers build intricate models to solve problems, but they can essentially be used as black boxes. What gets lost in there is whether we are actually using the hardware as effectively as we can,” says Siddharth Samsi, a research scientist in the LLSC. 

    To gain insight into this challenge, the LLSC has been collecting detailed data on TX-GAIA usage over the past year. More than a million user jobs later, the team has released the dataset open source to the computing community.

    Their goal is to empower computer scientists and data center operators to better understand avenues for data center optimization — an important task as processing needs continue to grow. They also see potential for leveraging AI in the data center itself, by using the data to develop models for predicting failure points, optimizing job scheduling, and improving energy efficiency. While cloud providers are actively working on optimizing their data centers, they do not often make their data or models available for the broader high-performance computing (HPC) community to leverage. The release of this dataset and associated code seeks to fill this space.

    “Data centers are changing. We have an explosion of hardware platforms, the types of workloads are evolving, and the types of people who are using data centers is changing,” says Vijay Gadepally, a senior researcher at the LLSC. “Until now, there hasn’t been a great way to analyze the impact to data centers. We see this research and dataset as a big step toward coming up with a principled approach to understanding how these variables interact with each other and then applying AI for insights and improvements.”

    Papers describing the dataset and potential applications have been accepted to a number of venues, including the IEEE International Symposium on High-Performance Computer Architecture, the IEEE International Parallel and Distributed Processing Symposium, the Annual Conference of the North American Chapter of the Association for Computational Linguistics, the IEEE High-Performance and Embedded Computing Conference, and International Conference for High Performance Computing, Networking, Storage and Analysis. 

    Workload classification

    Among the world’s TOP500 supercomputers, TX-GAIA combines traditional computing hardware (central processing units, or CPUs) with nearly 900 graphics processing unit (GPU) accelerators. These NVIDIA GPUs are specialized for deep learning, the class of AI that has given rise to speech recognition and computer vision.

    The dataset covers CPU, GPU, and memory usage by job; scheduling logs; and physical monitoring data. Compared to similar datasets, such as those from Google and Microsoft, the LLSC dataset offers “labeled data, a variety of known AI workloads, and more detailed time series data compared with prior datasets. To our knowledge, it’s one of the most comprehensive and fine-grained datasets available,” Gadepally says. 

    Notably, the team collected time-series data at an unprecedented level of detail: 100-millisecond intervals on every GPU and 10-second intervals on every CPU, as the machines processed more than 3,000 known deep-learning jobs. One of the first goals is to use this labeled dataset to characterize the workloads that different types of deep-learning jobs place on the system. This process would extract features that reveal differences in how the hardware processes natural language models versus image classification or materials design models, for example.   

    The team has now launched the MIT Datacenter Challenge to mobilize this research. The challenge invites researchers to use AI techniques to identify with 95 percent accuracy the type of job that was run, using their labeled time-series data as ground truth.

    Such insights could enable data centers to better match a user’s job request with the hardware best suited for it, potentially conserving energy and improving system performance. Classifying workloads could also allow operators to quickly notice discrepancies resulting from hardware failures, inefficient data access patterns, or unauthorized usage.

    Too many choices

    Today, the LLSC offers tools that let users submit their job and select the processors they want to use, “but it’s a lot of guesswork on the part of users,” Samsi says. “Somebody might want to use the latest GPU, but maybe their computation doesn’t actually need it and they could get just as impressive results on CPUs, or lower-powered machines.”

    Professor Devesh Tiwari at Northeastern University is working with the LLSC team to develop techniques that can help users match their workloads to appropriate hardware. Tiwari explains that the emergence of different types of AI accelerators, GPUs, and CPUs has left users suffering from too many choices. Without the right tools to take advantage of this heterogeneity, they are missing out on the benefits: better performance, lower costs, and greater productivity.

    “We are fixing this very capability gap — making users more productive and helping users do science better and faster without worrying about managing heterogeneous hardware,” says Tiwari. “My PhD student, Baolin Li, is building new capabilities and tools to help HPC users leverage heterogeneity near-optimally without user intervention, using techniques grounded in Bayesian optimization and other learning-based optimization methods. But, this is just the beginning. We are looking into ways to introduce heterogeneity in our data centers in a principled approach to help our users achieve the maximum advantage of heterogeneity autonomously and cost-effectively.”

    Workload classification is the first of many problems to be posed through the Datacenter Challenge. Others include developing AI techniques to predict job failures, conserve energy, or create job scheduling approaches that improve data center cooling efficiencies.

    Energy conservation 

    To mobilize research into greener computing, the team is also planning to release an environmental dataset of TX-GAIA operations, containing rack temperature, power consumption, and other relevant data.

    According to the researchers, huge opportunities exist to improve the power efficiency of HPC systems being used for AI processing. As one example, recent work in the LLSC determined that simple hardware tuning, such as limiting the amount of power an individual GPU can draw, could reduce the energy cost of training an AI model by 20 percent, with only modest increases in computing time. “This reduction translates to approximately an entire week’s worth of household energy for a mere three-hour time increase,” Gadepally says.

    They have also been developing techniques to predict model accuracy, so that users can quickly terminate experiments that are unlikely to yield meaningful results, saving energy. The Datacenter Challenge will share relevant data to enable researchers to explore other opportunities to conserve energy.

    The team expects that lessons learned from this research can be applied to the thousands of data centers operated by the U.S. Department of Defense. The U.S. Air Force is a sponsor of this work, which is being conducted under the USAF-MIT AI Accelerator.

    Other collaborators include researchers at MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). Professor Charles Leiserson’s Supertech Research Group is investigating performance-enhancing techniques for parallel computing, and research scientist Neil Thompson is designing studies on ways to nudge data center users toward climate-friendly behavior.

    Samsi presented this work at the inaugural AI for Datacenter Optimization (ADOPT’22) workshop last spring as part of the IEEE International Parallel and Distributed Processing Symposium. The workshop officially introduced their Datacenter Challenge to the HPC community.

    “We hope this research will allow us and others who run supercomputing centers to be more responsive to user needs while also reducing the energy consumption at the center level,” Samsi says. More

  • in

    New hardware offers faster computation for artificial intelligence, with much less energy

    As scientists push the boundaries of machine learning, the amount of time, energy, and money required to train increasingly complex neural network models is skyrocketing. A new area of artificial intelligence called analog deep learning promises faster computation with a fraction of the energy usage.

    Programmable resistors are the key building blocks in analog deep learning, just like transistors are the core elements for digital processors. By repeating arrays of programmable resistors in complex layers, researchers can create a network of analog artificial “neurons” and “synapses” that execute computations just like a digital neural network. This network can then be trained to achieve complex AI tasks like image recognition and natural language processing.

    A multidisciplinary team of MIT researchers set out to push the speed limits of a type of human-made analog synapse that they had previously developed. They utilized a practical inorganic material in the fabrication process that enables their devices to run 1 million times faster than previous versions, which is also about 1 million times faster than the synapses in the human brain.

    Moreover, this inorganic material also makes the resistor extremely energy-efficient. Unlike materials used in the earlier version of their device, the new material is compatible with silicon fabrication techniques. This change has enabled fabricating devices at the nanometer scale and could pave the way for integration into commercial computing hardware for deep-learning applications.

    “With that key insight, and the very powerful nanofabrication techniques we have at MIT.nano, we have been able to put these pieces together and demonstrate that these devices are intrinsically very fast and operate with reasonable voltages,” says senior author Jesús A. del Alamo, the Donner Professor in MIT’s Department of Electrical Engineering and Computer Science (EECS). “This work has really put these devices at a point where they now look really promising for future applications.”

    “The working mechanism of the device is electrochemical insertion of the smallest ion, the proton, into an insulating oxide to modulate its electronic conductivity. Because we are working with very thin devices, we could accelerate the motion of this ion by using a strong electric field, and push these ionic devices to the nanosecond operation regime,” explains senior author Bilge Yildiz, the Breene M. Kerr Professor in the departments of Nuclear Science and Engineering and Materials Science and Engineering.

    “The action potential in biological cells rises and falls with a timescale of milliseconds, since the voltage difference of about 0.1 volt is constrained by the stability of water,” says senior author Ju Li, the Battelle Energy Alliance Professor of Nuclear Science and Engineering and professor of materials science and engineering, “Here we apply up to 10 volts across a special solid glass film of nanoscale thickness that conducts protons, without permanently damaging it. And the stronger the field, the faster the ionic devices.”

    These programmable resistors vastly increase the speed at which a neural network is trained, while drastically reducing the cost and energy to perform that training. This could help scientists develop deep learning models much more quickly, which could then be applied in uses like self-driving cars, fraud detection, or medical image analysis.

    “Once you have an analog processor, you will no longer be training networks everyone else is working on. You will be training networks with unprecedented complexities that no one else can afford to, and therefore vastly outperform them all. In other words, this is not a faster car, this is a spacecraft,” adds lead author and MIT postdoc Murat Onen.

    Co-authors include Frances M. Ross, the Ellen Swallow Richards Professor in the Department of Materials Science and Engineering; postdocs Nicolas Emond and Baoming Wang; and Difei Zhang, an EECS graduate student. The research is published today in Science.

    Accelerating deep learning

    Analog deep learning is faster and more energy-efficient than its digital counterpart for two main reasons. “First, computation is performed in memory, so enormous loads of data are not transferred back and forth from memory to a processor.” Analog processors also conduct operations in parallel. If the matrix size expands, an analog processor doesn’t need more time to complete new operations because all computation occurs simultaneously.

    The key element of MIT’s new analog processor technology is known as a protonic programmable resistor. These resistors, which are measured in nanometers (one nanometer is one billionth of a meter), are arranged in an array, like a chess board.

    In the human brain, learning happens due to the strengthening and weakening of connections between neurons, called synapses. Deep neural networks have long adopted this strategy, where the network weights are programmed through training algorithms. In the case of this new processor, increasing and decreasing the electrical conductance of protonic resistors enables analog machine learning.

    The conductance is controlled by the movement of protons. To increase the conductance, more protons are pushed into a channel in the resistor, while to decrease conductance protons are taken out. This is accomplished using an electrolyte (similar to that of a battery) that conducts protons but blocks electrons.

    To develop a super-fast and highly energy efficient programmable protonic resistor, the researchers looked to different materials for the electrolyte. While other devices used organic compounds, Onen focused on inorganic phosphosilicate glass (PSG).

    PSG is basically silicon dioxide, which is the powdery desiccant material found in tiny bags that come in the box with new furniture to remove moisture. It is studied as a proton conductor under humidified conditions for fuel cells. It is also the most well-known oxide used in silicon processing. To make PSG, a tiny bit of phosphorus is added to the silicon to give it special characteristics for proton conduction.

    Onen hypothesized that an optimized PSG could have a high proton conductivity at room temperature without the need for water, which would make it an ideal solid electrolyte for this application. He was right.

    Surprising speed

    PSG enables ultrafast proton movement because it contains a multitude of nanometer-sized pores whose surfaces provide paths for proton diffusion. It can also withstand very strong, pulsed electric fields. This is critical, Onen explains, because applying more voltage to the device enables protons to move at blinding speeds.

    “The speed certainly was surprising. Normally, we would not apply such extreme fields across devices, in order to not turn them into ash. But instead, protons ended up shuttling at immense speeds across the device stack, specifically a million times faster compared to what we had before. And this movement doesn’t damage anything, thanks to the small size and low mass of protons. It is almost like teleporting,” he says.

    “The nanosecond timescale means we are close to the ballistic or even quantum tunneling regime for the proton, under such an extreme field,” adds Li.

    Because the protons don’t damage the material, the resistor can run for millions of cycles without breaking down. This new electrolyte enabled a programmable protonic resistor that is a million times faster than their previous device and can operate effectively at room temperature, which is important for incorporating it into computing hardware.

    Thanks to the insulating properties of PSG, almost no electric current passes through the material as protons move. This makes the device extremely energy efficient, Onen adds.

    Now that they have demonstrated the effectiveness of these programmable resistors, the researchers plan to reengineer them for high-volume manufacturing, says del Alamo. Then they can study the properties of resistor arrays and scale them up so they can be embedded into systems.

    At the same time, they plan to study the materials to remove bottlenecks that limit the voltage that is required to efficiently transfer the protons to, through, and from the electrolyte.

    “Another exciting direction that these ionic devices can enable is energy-efficient hardware to emulate the neural circuits and synaptic plasticity rules that are deduced in neuroscience, beyond analog deep neural networks. We have already started such a collaboration with neuroscience, supported by the MIT Quest for Intelligence,” adds Yildiz.

    “The collaboration that we have is going to be essential to innovate in the future. The path forward is still going to be very challenging, but at the same time it is very exciting,” del Alamo says.

    “Intercalation reactions such as those found in lithium-ion batteries have been explored extensively for memory devices. This work demonstrates that proton-based memory devices deliver impressive and surprising switching speed and endurance,” says William Chueh, associate professor of materials science and engineering at Stanford University, who was not involved with this research. “It lays the foundation for a new class of memory devices for powering deep learning algorithms.”

    “This work demonstrates a significant breakthrough in biologically inspired resistive-memory devices. These all-solid-state protonic devices are based on exquisite atomic-scale control of protons, similar to biological synapses but at orders of magnitude faster rates,” says Elizabeth Dickey, the Teddy & Wilton Hawkins Distinguished Professor and head of the Department of Materials Science and Engineering at Carnegie Mellon University, who was not involved with this work. “I commend the interdisciplinary MIT team for this exciting development, which will enable future-generation computational devices.”

    This research is funded, in part, by the MIT-IBM Watson AI Lab. More