Designing cleaner vehicles

Adi Mehrotra knew that his time at MIT wasn’t up yet when he finished his undergraduate degree in 2022. During his first four years at the Institute, he was a critical member of the Solar Electric Vehicle Team (SEVT) and eventually led the group to victory in a five-day, 900-mile race. Later, he translated the skills he learned from SEVT to a summer internship in Ghana with the startup Moving Health, where he worked on low-cost ambulances that could transport patients from remote villages to medical care, without relying on gasoline. But there were still more projects he wanted to tackle.

Now, as a second-year master’s degree student in mechanical engineering, Mehrotra has channeled his energy into two arenas: designing clean energy vehicles and enhancing mechanical engineering education at MIT. For the former, he has taken the helm of the MIT Electric Vehicle Team, a student-led research team that is probing the future of transportation by designing a hydrogen-powered motorcycle. And for his master’s thesis research, Mehrotra is building a new mechatronics curriculum, an interdisciplinary course at the intersection of mechanical and electrical engineering.

Building in the basement

Mehrotra cannot remember a time in his life when science did not consume his attention. He credits his parents with fostering this interest, by encouraging scientific thinking with subscriptions to Ask magazine and National Geographic, and lots of LEGO play from a very young age.

As a high school student in East Brunswick, New Jersey, his passion for building blossomed. Mehrotra says, “I built a lot of my own projects, in the basement, mostly relying on wood or metal.” He also co-founded his school’s FIRST Robotics team chapter. He says, “It’s an experience I loved and cherish to this day, as it gave me a lot of hands-on building experience and allowed to explore creative ways to solve exciting problems.”

However, launching the team was much more challenging than Mehrotra expected. At the time, his high school didn’t want to assume any liability associated with the chapter’s activities, so the team operated out of his basement. “Most of our robots were built with a single drill, a Dremel, and a single drill press,” he says.

Given those constraints, Mehrotra was pleased with how well the team performed in the annual regional competition. His parents were very supportive, despite the late nights and loud noises coming from their basement that sometimes compromised their sleep. Mehrotra says, “I think they were very glad that I had something so meaningful and fulfilling to work on.” Today, the team that he started is still going strong, which fills him with pride.

The human side of MIT

“My mom always knew she wanted me to go to MIT,” Mehrotra says with a laugh. But he wanted to see MIT for himself. Campus Preview Weekend (CPW), when admitted students visit the Institute, gave him that opportunity. That April weekend, he observed East Campus students building a rollercoaster out of wood — a time-honored tradition. “Even as a pre-first-year, they told me to take a drill and help them build. I knew of MIT as one of the best mechanical engineering departments in the world. CPW sealed MIT as a very human place, as well,” he says.

As an undergraduate, Mehrotra majored in electrical engineering and computer science. However, he says, “I realized pretty quickly that I don’t like software design, and took as many mechanical engineering classes as I possibly could.” He also joined professor of mechanical engineering Sangbae Kim’s Biomimetic Robotics lab as a junior, which further cemented his passion for mechanical design.

Mehrotra became especially drawn to vehicle design, an interest that he cultivated in a variety of ways. He joined the Solar Electric Vehicle Team his first year and remained involved for three years. The team designs and builds a solar-powered vehicle and races it in international competitions. Mehrotra started out as a member of the aero team, working on the aerodynamics of the car; over time, he rose through the ranks to become the team captain, leading his peers to victory in the 2021 race from Independence, Missouri, to Las Vegas, New Mexico. In addition to learning about mechanical design, he also learned about life. He says, “The people on the team remain some of my best friends. The older people in solar car were the best personal and engineering mentors. They taught me how to lead a team and treat people well.”

Mehrotra’s enthusiasm for design also flourished through MIT D-Lab, an initiative that designs solutions for use in the developing world. He says, “There are a lot of classes at MIT that have taught me a lot of things, of course. However, in D-Lab, I walked in with one assumption about good ways to make the world a better place and they kind of flipped that on its head. [D-Lab] approaches problem-solving from this local perspective that if you can help one person very well, that is a bigger success than helping 100 people poorly.”

His experience in D-Lab’s course 2.729 (Design for Scale) led him to a pivotal mentor, Emily Young ’18, the founder of the startup Moving Health, based in Ghana. Mehrotra spent the summer of 2022 there as an intern, building tricycle ambulances to connect rural regions of Ghana to urban medical centers.

A new outlook

Mehrotra returned from Ghana to begin grad school armed with a fresh perspective on academics. “In the past, sometimes I was afraid of taking certain classes, or doing certain activities, because I didn’t feel like I had the right background, but I’ve thrown that fear away in graduate school,” he says.

In that spirit, Mehrotra took a risk in his choice of master’s thesis project: He decided to develop a new curriculum for mechatronics at MIT, even though he had little prior experience in curriculum development. He joined Professor Sangbae Kim’s lab again, drawn by Kim’s dedication to mentoring and teaching, and he served alongside Kim as a TA for the legendary course 2.007 (Design and Manufacturing I).

“Working with someone who shares similar teaching philosophies to me is really cool to learn from, and you also feel like you are making an impact through teaching, especially while working with Sangbae,” Mehrotra says. His master’s thesis will focus on addressing the limitations of the mechatronics curriculum. “We think the current curriculum is not adequate to prepare our students for careers in industry and academia. So, we’re looking to incorporate the psychology of learning into building a better curriculum,” he explains.

In keeping with his affinity for vehicles, motorcycles seem to be another theme for Mehrotra. Inspired by his time in Ghana, where motorcycles are among the most common modes of transportation, he bought a vintage model for himself, a 1974 Honda CB360 — and he braves Boston traffic on a regular basis on his bike. In addition, he has devoted much of his time to the Electric Vehicle Team, where he is working on a hydrogen-powered motorcycle. Despite the team’s moniker, Mehrotra says, “We as a team are not saying that … we should get rid of all battery-powered cars immediately, but we would like to try a proof-of-concept on our own.” The team’s motorcycle is fondly named Toothless, a nod to the dragon in the Dreamworks movie “How to Train Your Dragon.”

In his spare time, Mehrotra also dabbles in cinematography and film. He is currently working on a documentary in collaboration with Moving Health. “I met so many amazing people in Ghana and we want to be able to tell their amazing stories, but we also want to change perceptions of people who live in underserved communities. We often do not talk about them in fair and meaningful ways, but instead just assume that they are helpless,” he says.

Once he finishes his master’s degree, Mehrotra plans to pursue a PhD under Professor Alex Slocum, the Walter M. May and A. Hazel May Chair in Emerging Technologies, focusing on hydrogen energy systems. Influenced by his time in Ghana, Mehrotra has realized that he wants to pursue research that could impact the developing world. He says, “Climate change disproportionally affects people who live in underserved communities around the world, despite the fact that most of climate change’s causes originate from western nations. Solving the energy crisis has implications to many of the United Nations Sustainable Development Goals, and social impacts far beyond just mitigating climate change.”

de MIT News https://ift.tt/vMaPG8O

Steven Barrett named head of the Department of Aeronautics and Astronautics

Steven Barrett, the H.N. Slater Professor of Aeronautics and Astronautics, has been named the new head of the Department of Aeronautics and Astronautics (AeroAstro), effective Nov. 1.

“Professor Barrett is incredibly well-suited to serve as leader of the Department of Aeronautics and Astronautics. Having served as associate department head since 2021 and interim department head for the past five months, he has demonstrated a commitment to the AeroAstro community,” says Anantha Chandrakasan, dean of the MIT School of Engineering and the Vannevar Bush Professor of Electrical Engineering and Computer Science. “He is a dedicated educator and brilliant researcher. I am looking forward to working closely with him in this new role.”

Barrett joined the AeroAstro departmental leadership in July 2021 as associate department head. In May, Barrett was named interim department head after Daniel Hastings, the Cecil and Ida Green Education Professor of Aeronautics and Astronautics, announced he was stepping down to serve as interim Institute community and equity officer and interim associate provost for faculty development.

As director of the MIT Laboratory for Aviation and the Environment, Barrett and his team study the impact aviation has on the environment. He has developed a number of solutions to mitigate the impact aviation has on air quality, climate, and noise pollution. The overall goal of his research is to help develop technologies that eliminate the environmental impact of aviation.

As associate department head, Barrett focused on AeroAstro’s graduate and undergraduate education. In addition to developing programs that aim to increase funding for graduate students, Barrett contributed to the creation of the AeroAstro HBCU Partnership Program. He also co-led the development of the department’s digital education strategy, which aims to bring aerospace education to more people across the world.

Since joining MIT’s faculty in 2010, Barrett has been involved in various programs across MIT. He has participated in the MIT Joint Program on the Science and Policy of Global Change, the SMART Center for Environmental Sensing and Modeling, the MIT Environmental Solutions Initiative, the MIT Energy Initiative, and the MIT Climate and Sustainability Consortium.

Barrett’s research has been selected for a grant under the NASA Innovative Advanced Concepts Program. He is also the recipient of a Bose Research Grant. His work on the first-ever plane with no moving propulsion parts was honored as one of the 10 Breakthroughs of 2018 by Physics World. From 2011 to 2015, Barrett served as associate director of the Partnership for Air Transportation Noise and Emissions Reduction, which included 50 participants across academia, industry, and government.

As an undergraduate student at the University of Cambridge, he studied for a year at MIT through an exchange program. He received his bachelor’s degree, master’s degree, and PhD from the University of Cambridge, where he served as a faculty member for one year prior to joining MIT’s faculty in 2010.

Barrett succeeds Daniel Hastings, who has led AeroAstro since January 2019.

“I am incredibly grateful to Professor Hastings for his tremendous leadership, not only as head of AeroAstro but as associate dean for diversity, equity, and inclusion in the School of Engineering. He has been a trusted advisor for many years. His contributions will have an impact on AeroAstro and the wider engineering community for many years to come,” adds Chandrakasan.

de MIT News https://ift.tt/pPXfB6A

In a surprising finding, light can make water evaporate without heat

Evaporation is happening all around us all the time, from the sweat cooling our bodies to the dew burning off in the morning sun. But science’s understanding of this ubiquitous process may have been missing a piece all this time.

In recent years, some researchers have been puzzled upon finding that water in their experiments, which was held in a sponge-like material known as a hydrogel, was evaporating at a higher rate than could be explained by the amount of heat, or thermal energy, that the water was receiving. And the excess has been significant — a doubling, or even a tripling or more, of the theoretical maximum rate.

After carrying out a series of new experiments and simulations, and reexamining some of the results from various groups that claimed to have exceeded the thermal limit, a team of researchers at MIT has reached a startling conclusion: Under certain conditions, at the interface where water meets air, light can directly bring about evaporation without the need for heat, and it actually does so even more efficiently than heat. In these experiments, the water was held in a hydrogel material, but the researchers suggest that the phenomenon may occur under other conditions as well.

The findings are published this week in a paper in PNAS, by MIT postdoc Yaodong Tu, professor of mechanical engineering Gang Chen, and four others.

The phenomenon might play a role in the formation and evolution of fog and clouds, and thus would be important to incorporate into climate models to improve their accuracy, the researchers say. And it might play an important part in many industrial processes such as solar-powered desalination of water, perhaps enabling alternatives to the step of converting sunlight to heat first.

The new findings come as a surprise because water itself does not absorb light to any significant degree. That’s why you can see clearly through many feet of clean water to the surface below. So, when the team initially began exploring the process of solar evaporation for desalination, they first put particles of a black, light-absorbing material in a container of water to help convert the sunlight to heat.

Then, the team came across the work of another group that had achieved an evaporation rate double the thermal limit — which is the highest possible amount of evaporation that can take place for a given input of heat, based on basic physical principles such as the conservation of energy. It was in these experiments that the water was bound up in a hydrogel. Although they were initially skeptical, Chen and Tu starting their own experiments with hydrogels, including a piece of the material from the other group. “We tested it under our solar simulator, and it worked,” confirming the unusually high evaporation rate, Chen says. “So, we believed them now.” Chen and Tu then began making and testing their own hydrogels.

They began to suspect that the excess evaporation was being caused by the light itself —that photons of light were actually knocking bundles of water molecules loose from the water’s surface. This effect would only take place right at the boundary layer between water and air, at the surface of the hydrogel material — and perhaps also on the sea surface or the surfaces of droplets in clouds or fog.

In the lab, they monitored the surface of a hydrogel, a JELL-O-like matrix consisting mostly of water bound by a sponge-like lattice of thin membranes. They measured its responses to simulated sunlight with precisely controlled wavelengths.

The researchers subjected the water surface to different colors of light in sequence and measured the evaporation rate. They did this by placing a container of water-laden hydrogel on a scale and directly measuring the amount of mass lost to evaporation, as well as monitoring the temperature above the hydrogel surface. The lights were shielded to prevent them from introducing extra heat. The researchers found that the effect varied with color and peaked at a particular wavelength of green light. Such a color dependence has no relation to heat, and so supports the idea that it is the light itself that is causing at least some of the evaporation.

The researchers tried to duplicate the observed evaporation rate with the same setup but using electricity to heat the material, and no light. Even though the thermal input was the same as in the other test, the amount of water that evaporated never exceeded the thermal limit. However, it did so when the simulated sunlight was on, confirming that light was the cause of the extra evaporation.

Though water itself does not absorb much light, and neither does the hydrogel material itself, when the two combine they become strong absorbers, Chen says. That allows the material to harness the energy of the solar photons efficiently and exceed the thermal limit, without the need for any dark dyes for absorption.

Having discovered this effect, which they have dubbed the photomolecular effect, the researchers are now working on how to apply it to real-world needs. They have a grant from the Abdul Latif Jameel Water and Food Systems Lab to study the use of this phenomenon to improve the efficiency of solar-powered desalination systems, and a Bose Grant to explore the phenomenon’s effects on climate change modeling.

Tu explains that in standard desalination processes, “it normally has two steps: First we evaporate the water into vapor, and then we need to condense the vapor to liquify it into fresh water.” With this discovery, he says, potentially “we can achieve high efficiency on the evaporation side.” The process also could turn out to have applications in processes that require drying a material.

Chen says that in principle, he thinks it may be possible to increase the limit of water produced by solar desalination, which is currently 1.5 kilograms per square meter, by as much as three- or fourfold using this light-based approach. “This could potentially really lead to cheap desalination,” he says.

Tu adds that this phenomenon could potentially also be leveraged in evaporative cooling processes, using the phase change to provide a highly efficient solar cooling system.

Meanwhile, the researchers are also working closely with other groups who are attempting to replicate the findings, hoping to overcome skepticism that has faced the unexpected findings and the hypothesis being advanced to explain them.

The research team also included Jiawei Zhou, Shaoting Lin, Mohammed Alshrah, and Xuanhe Zhao, all in MIT’s Department of Mechanical Engineering.

de MIT News https://ift.tt/yMAdoZL

Engineers develop an efficient process to make fuel from carbon dioxide

The search is on worldwide to find ways to extract carbon dioxide from the air or from power plant exhaust and then make it into something useful. One of the more promising ideas is to make it into a stable fuel that can replace fossil fuels in some applications. But most such conversion processes have had problems with low carbon efficiency, or they produce fuels that can be hard to handle, toxic, or flammable.

Now, researchers at MIT and Harvard University have developed an efficient process that can convert carbon dioxide into formate, a liquid or solid material that can be used like hydrogen or methanol to power a fuel cell and generate electricity. Potassium or sodium formate, already produced at industrial scales and commonly used as a de-icer for roads and sidewalks, is nontoxic, nonflammable, easy to store and transport, and can remain stable in ordinary steel tanks to be used months, or even years, after its production.

The new process, developed by MIT doctoral students Zhen Zhang, Zhichu Ren, and Alexander H. Quinn; Harvard University doctoral student Dawei Xi; and MIT Professor Ju Li, is described this week in an open-access paper in Cell Reports Physical Science. The whole process — including capture and electrochemical conversion of the gas to a solid formate powder, which is then used in a fuel cell to produce electricity — was demonstrated at a small, laboratory scale. However, the researchers expect it to be scalable so that it could provide emissions-free heat and power to individual homes and even be used in industrial or grid-scale applications.

Other approaches to converting carbon dioxide into fuel, Li explains, usually involve a two-stage process: First the gas is chemically captured and turned into a solid form as calcium carbonate, then later that material is heated to drive off the carbon dioxide and convert it to a fuel feedstock such as carbon monoxide. That second step has very low efficiency, typically converting less than 20 percent of the gaseous carbon dioxide into the desired product, Li says.

By contrast, the new process achieves a conversion of well over 90 percent and eliminates the need for the inefficient heating step by first converting the carbon dioxide into an intermediate form, liquid metal bicarbonate. That liquid is then electrochemically converted into liquid potassium or sodium formate in an electrolyzer that uses low-carbon electricity, e.g. nuclear, wind, or solar power. The highly concentrated liquid potassium or sodium formate solution produced can then be dried, for example by solar evaporation, to produce a solid powder that is highly stable and can be stored in ordinary steel tanks for up to years or even decades, Li says.

Several steps of optimization developed by the team made all the difference in changing an inefficient chemical-conversion process into a practical solution, says Li, who holds joint appointments in the departments of Nuclear Science and Engineering and of Materials Science and Engineering.

The process of carbon capture and conversion involves first an alkaline solution-based capture that concentrates carbon dioxide, either from concentrated streams such as from power plant emissions or from very low-concentration sources, even open air, into the form of a liquid metal-bicarbonate solution. Then, through the use of a cation-exchange membrane electrolyzer, this bicarbonate is electrochemically converted into solid formate crystals with a carbon efficiency of greater than 96 percent, as confirmed in the team’s lab-scale experiments.

These crystals have an indefinite shelf life, remaining so stable that they could be stored for years, or even decades, with little or no loss. By comparison, even the best available practical hydrogen storage tanks allow the gas to leak out at a rate of about 1 percent per day, precluding any uses that would require year-long storage, Li says. Methanol, another widely explored alternative for converting carbon dioxide into a fuel usable in fuel cells, is a toxic substance that cannot easily be adapted to use in situations where leakage could pose a health hazard. Formate, on the other hand, is widely used and considered benign, according to national safety standards.

Several improvements account for the greatly improved efficiency of this process. First, a careful design of the membrane materials and their configuration overcomes a problem that previous attempts at such a system have encountered, where a buildup of certain chemical byproducts changes the pH, causing the system to steadily lose efficiency over time. “Traditionally, it is difficult to achieve long-term, stable, continuous conversion of the feedstocks,” Zhang says. “The key to our system is to achieve a pH balance for steady-state conversion.”

To achieve that, the researchers carried out thermodynamic modeling to design the new process so that it is chemically balanced and the pH remains at a steady state with no shift in acidity over time. It can therefore continue operating efficiently over long periods. In their tests, the system ran for over 200 hours with no significant decrease in output. The whole process can be done at ambient temperatures and relatively low pressures (about five times atmospheric pressure).

Another issue was that unwanted side reactions produced other chemical products that were not useful, but the team figured out a way to prevent these side reactions by the introduction of an extra “buffer” layer of bicarbonate-enriched fiberglass wool that blocked these reactions.

The team also built a fuel cell specifically optimized for the use of this formate fuel to produce electricity. The stored formate particles are simply dissolved in water and pumped into the fuel cell as needed. Although the solid fuel is much heavier than pure hydrogen, when the weight and volume of the high-pressure gas tanks needed to store hydrogen is considered, the end result is an electricity output near parity for a given storage volume, Li says.

The formate fuel can potentially be adapted for anything from home-sized units to large scale industrial uses or grid-scale storage systems, the researchers say. Initial household applications might involve an electrolyzer unit about the size of a refrigerator to capture and convert the carbon dioxide into formate, which could be stored in an underground or rooftop tank. Then, when needed, the powdered solid would be mixed with water and fed into a fuel cell to provide power and heat. “This is for community or household demonstrations,” Zhang says, “but we believe that also in the future it may be good for factories or the grid.”

“The formate economy is an intriguing concept because metal formate salts are very benign and stable, and a compelling energy carrier,” says Ted Sargent, a professor of chemistry and of electrical and computer engineering at Northwestern University, who was not associated with this work. “The authors have demonstrated enhanced efficiency in liquid-to-liquid conversion from bicarbonate feedstock to formate, and have demonstrated these fuels can be used later to produce electricity,” he says.

The work was supported by the U.S. Department of Energy Office of Science.

de MIT News https://ift.tt/nK6UdHX

The brain may learn about the world the same way some computational models do

To make our way through the world, our brain must develop an intuitive understanding of the physical world around us, which we then use to interpret sensory information coming into the brain.

How does the brain develop that intuitive understanding? Many scientists believe that it may use a process similar to what’s known as “self-supervised learning.” This type of machine learning, originally developed as a way to create more efficient models for computer vision, allows computational models to learn about visual scenes based solely on the similarities and differences between them, with no labels or other information.

A pair of studies from researchers at the K. Lisa Yang Integrative Computational Neuroscience (ICoN) Center at MIT offers new evidence supporting this hypothesis. The researchers found that when they trained models known as neural networks using a particular type of self-supervised learning, the resulting models generated activity patterns very similar to those seen in the brains of animals that were performing the same tasks as the models.

The findings suggest that these models are able to learn representations of the physical world that they can use to make accurate predictions about what will happen in that world, and that the mammalian brain may be using the same strategy, the researchers say.

“The theme of our work is that AI designed to help build better robots ends up also being a framework to better understand the brain more generally,” says Aran Nayebi, a postdoc in the ICoN Center. “We can’t say if it’s the whole brain yet, but across scales and disparate brain areas, our results seem to be suggestive of an organizing principle.”

Nayebi is the lead author of one of the studies, co-authored with Rishi Rajalingham, a former MIT postdoc now at Meta Reality Labs, and senior authors Mehrdad Jazayeri, an associate professor of brain and cognitive sciences and a member of the McGovern Institute for Brain Research; and Robert Yang, an assistant professor of brain and cognitive sciences and an associate member of the McGovern Institute. Ila Fiete, director of the ICoN Center, a professor of brain and cognitive sciences, and an associate member of the McGovern Institute, is the senior author of the other study, which was co-led by Mikail Khona, an MIT graduate student, and Rylan Schaeffer, a former senior research associate at MIT.

Both studies will be presented at the 2023 Conference on Neural Information Processing Systems (NeurIPS) in December.

Modeling the physical world

Early models of computer vision mainly relied on supervised learning. Using this approach, models are trained to classify images that are each labeled with a name — cat, car, etc. The resulting models work well, but this type of training requires a great deal of human-labeled data.

To create a more efficient alternative, in recent years researchers have turned to models built through a technique known as contrastive self-supervised learning. This type of learning allows an algorithm to learn to classify objects based on how similar they are to each other, with no external labels provided.

“This is a very powerful method because you can now leverage very large modern data sets, especially videos, and really unlock their potential,” Nayebi says. “A lot of the modern AI that you see now, especially in the last couple years with ChatGPT and GPT-4, is a result of training a self-supervised objective function on a large-scale dataset to obtain a very flexible representation.”

These types of models, also called neural networks, consist of thousands or millions of processing units connected to each other. Each node has connections of varying strengths to other nodes in the network. As the network analyzes huge amounts of data, the strengths of those connections change as the network learns to perform the desired task.

As the model performs a particular task, the activity patterns of different units within the network can be measured. Each unit’s activity can be represented as a firing pattern, similar to the firing patterns of neurons in the brain. Previous work from Nayebi and others has shown that self-supervised models of vision generate activity similar to that seen in the visual processing system of mammalian brains.

In both of the new NeurIPS studies, the researchers set out to explore whether self-supervised computational models of other cognitive functions might also show similarities to the mammalian brain. In the study led by Nayebi, the researchers trained self-supervised models to predict the future state of their environment across hundreds of thousands of naturalistic videos depicting everyday scenarios.    

“For the last decade or so, the dominant method to build neural network models in cognitive neuroscience is to train these networks on individual cognitive tasks. But models trained this way rarely generalize to other tasks,” Yang says. “Here we test whether we can build models for some aspect of cognition by first training on naturalistic data using self-supervised learning, then evaluating in lab settings.”

Once the model was trained, the researchers had it generalize to a task they call “Mental-Pong.” This is similar to the video game Pong, where a player moves a paddle to hit a ball traveling across the screen. In the Mental-Pong version, the ball disappears shortly before hitting the paddle, so the player has to estimate its trajectory in order to hit the ball.

The researchers found that the model was able to track the hidden ball’s trajectory with accuracy similar to that of neurons in the mammalian brain, which had been shown in a previous study by Rajalingham and Jazayeri to simulate its trajectory — a cognitive phenomenon known as “mental simulation.” Furthermore, the neural activation patterns seen within the model were similar to those seen in the brains of animals as they played the game — specifically, in a part of the brain called the dorsomedial frontal cortex. No other class of computational model has been able to match the biological data as closely as this one, the researchers say.

“There are many efforts in the machine learning community to create artificial intelligence,” Jazayeri says. “The relevance of these models to neurobiology hinges on their ability to additionally capture the inner workings of the brain. The fact that Aran’s model predicts neural data is really important as it suggests that we may be getting closer to building artificial systems that emulate natural intelligence.”

Navigating the world

The study led by Khona, Schaeffer, and Fiete focused on a type of specialized neurons known as grid cells. These cells, located in the entorhinal cortex, help animals to navigate, working together with place cells located in the hippocampus.

While place cells fire whenever an animal is in a specific location, grid cells fire only when the animal is at one of the vertices of a triangular lattice. Groups of grid cells create overlapping lattices of different sizes, which allows them to encode a large number of positions using a relatively small number of cells.

In recent studies, researchers have trained supervised neural networks to mimic grid cell function by predicting an animal’s next location based on its starting point and velocity, a task known as path integration. However, these models hinged on access to privileged information about absolute space at all times — information that the animal does not have.                               

Inspired by the striking coding properties of the multiperiodic grid-cell code for space, the MIT team trained a contrastive self-supervised model to both perform this same path integration task and represent space efficiently while doing so. For the training data, they used sequences of velocity inputs. The model learned to distinguish positions based on whether they were similar or different — nearby positions generated similar codes, but further positions generated more different codes.    

“It’s similar to training models on images, where if two images are both heads of cats, their codes should be similar, but if one is the head of a cat and one is a truck, then you want their codes to repel,” Khona says. “We’re taking that same idea but applying it to spatial trajectories.”

Once the model was trained, the researchers found that the activation patterns of the nodes within the model formed several lattice patterns with different periods, very similar to those formed by grid cells in the brain.

“What excites me about this work is that it makes connections between mathematical work on the striking information-theoretic properties of the grid cell code and the computation of path integration,” Fiete says. “While the mathematical work was analytic — what properties does the grid cell code possess? — the approach of optimizing coding efficiency through self-supervised learning and obtaining grid-like tuning is synthetic: It shows what properties might be necessary and sufficient to explain why the brain has grid cells.”

The research was funded by the K. Lisa Yang ICoN Center, the National Institutes of Health, the Simons Foundation, the McKnight Foundation, the McGovern Institute, and the Helen Hay Whitney Foundation.

de MIT News https://ift.tt/lXpQDfG

3 Questions: A roadmap toward circularity in the footwear industry

In March 2022, representatives of global footwear brands gathered with sustainability experts and other academic researchers on the MIT campus. The mission? To kick-start a discussion on addressing the waste produced by the footwear industry. An MIT research project arising from the Circular Shoe Systems Summit has now resulted in a white paper titled the Footwear Manifesto. The report is co-authored by Yuly Fuentes-Medel, as MIT’s Fabric Innovation Hub manager and now program director of textiles at MIT Climate Grand Challenges; Leslie Yan ’22, who majored in mechanical engineering and design; and collaborator Karo Buttler, who graduated last year from the Fashion Institute of Technology in New York.

In the months following the summit, the team conducted in-depth interviews and a broad survey of the industry to bring the Footwear Manifesto to life. The report aims to provide a global overview of current barriers and challenges to achieving a more sustainable and circular footwear industry — that is, one that incorporates reuse, recovery, recycling, and regeneration to cut down on waste. In a conversation prepared for MIT News, Fuentes-Medel, Yan, and Buttler reflect on their findings. Fuentes-Medel also suggests an opportunity for pre-competitive collaborative research to move the needle on climate action and circularity in the footwear industry.

Q: What are the main findings in this report?

Fuentes-Medel: In building this report, we spoke to more than 15 footwear companies and numerous external stakeholders about the status quo of sustainability in the sector. The document identifies opportunities to steer the industry towards environmentally sound solutions across several aspects of the business: materials management, post-consumer infrastructure, consumer behavior, and implementation of circular business models. Our survey, interviews, and analysis identified several key findings:

• Today, the complexity of shoes — the design complexity, manufacturing process, and temporal use — is the biggest challenge to circularity at scale.
• Companies have the will to change, but there is little strategic alignment on circularity across the footwear industry, and no common means of measuring progress.
• Collaboration is vital for scaling circular systems — not just within the brands, but with a multitude of stakeholders including suppliers, infrastructure investors, government, academia, and entrepreneurs.
• Building a circular dynamic system will require diverse solutions that open opportunities across sectors and communities.

Q: Why shoes? What are the unique challenges facing the footwear industry?

Fuentes-Medel: Everyone wears shoes. Whether it’s for protection, comfort and support, performance, or self-expression, footwear plays a major role in getting us where we need to go. Most people, however, have no idea what happens to their shoes when they’ve finished using them. The linear take-make-waste model of the footwear industry is deeply unsustainable from a social, economic, and environmental perspective.

A fundamental barrier to achieving a circular shoe system lies with the complexity of shoes themselves, as reflected in their design, material utilization, and manufacturing processes. For example, the intricate multi-material construction of shoes, consisting of dozens of individual components, makes it nearly impossible at present to recycle and reintegrate used shoes back into the supply chain.

However, the challenges faced by the industry extend beyond the shoe itself. Achieving greater sustainability in footwear necessitates a dedicated commitment to resource stewardship. A shoe designed to be recyclable or biodegradable cannot genuinely be considered "circular" unless there is a robust logistical infrastructure that oversees the collection, sorting, processing, and disassembly of the shoe at the end of its life. This infrastructure is essential for enabling the reuse, recycling, or composting of the finished product. Achieving progress in sustainable manufacturing and consumption requires extensive coordination among a wide network of suppliers, buyers, and material management entities. If such system existed, it could create new jobs and opportunities for value creation.

I have continued to collaborate with a group of participants from the summit to advance that vision by establishing — outside of MIT — a consortium of brands and industry partners that will be known as The Footwear Collective, which will exist to tackle some of these challenges related to circularity under the sponsorship of the nonprofit EarthDNA. Embracing the circular economy will require an inclusive shift in behavior from mere sharing to active participation. All footwear brands are invited to participate.

Q: Leslie, Karo, as students working on this project, what were some of your biggest takeaways?

Yan: Coming from a design and engineering background, I was initially drawn to investigating the sustainability challenges inherent in the way shoes are constructed and produced. However, it became evident that the environmental burden of the industry is also a product of barriers posed by the complex system in which shoes are made, sold, and consumed. Still, this complexity opens up multiple pathways towards greener practices, ranging from materials to consumer behavior — each of which we outline in our report. It was truly exciting to see the determination and enthusiasm of many in the industry towards taking meaningful steps towards a more sustainable future for footwear.

Buttler: Our goal for this report was to curate and centralize already-existing information for the footwear industry. During that process, we recognized the necessity of collaboration among brands and various industries. As a young designer, the excitement for more sustainable practices was very inspiring, and I can’t wait to see where the footwear industry will be in 5-10 years.

de MIT News https://ift.tt/xstJaLG

MIT’s Justin Yu wins Classic Tetris World Championship

If you have 59 seconds to devote to pure joy, you won't regret watching this video clip of Justin “Fractal” Yu, an MIT junior who, on Oct. 15, became the top classic Tetris player in the world by winning the 2023 Classic Tetris World Championships.

The computer science and engineering major from Dallas, Texas, plans to double major in math with a minor in music technology. Despite his busy schedule, he still finds time to play the cello in MIT’s Videogame Orchestra, and to practice one very classic, very challenging Nintendo Entertainment System console game enough to become the single greatest player on the planet. Here, he describes his early experiences with classic Tetris, his training regimen, what he did to celebrate his big win, his vision for the future, and more.

Q:Tell us about when you first encountered classic Tetris. 

A: The first time I heard of it was about seven years ago. I watched other players and championships on YouTube for about two years before starting to play myself when a friend got really into it. I wanted to beat his score, discovered it was really hard, and it all snowballed from there. 

Q: How does classic Tetris differ from the Tetris you might find on your phone? Tell us about the physical game controller that we can see you manipulating in the video. 



A: Any Tetris game that you’ll find nowadays has to obey a certain standard mandated by the Tetris company. Each piece has to have a specific color, you have to see three pieces in advance, you have to have a piece you can swap out at any time. 

But classic Tetris, which came out in 1989 for the Nintendo Entertainment System (NES), only lets you see one piece at a time. You don’t have the hold option to swap out a piece, and most importantly, you aren’t able to customize how pieces move left and right. We had to find a way around these obstacles, and this gets to the controller thing. In all the modern versions of the game, you can move pieces from side to side instantly just by holding a button — but that mechanic doesn’t exist in classic Tetris. So we had to work around this, and perfected this technique where we can mash buttons as fast as humanly possible, called “rolling.” You have to use all your fingers to push the controller into your hand. Since the game’s frame rate is 60 hertz, the maximum possible input is 30 hertz, and with this technique we’re able to get it. 

The rules in competition are simple. You play on your own setup and get your own score. In the interest of fairness, we play with the same pieces on each side, so each player gets approximately the same amount of luck. 



Q: What were your early impressions of the game’s difficulty and technique, and how did they change as you played it more?



A: The learning curve is very steep and punishing, because mistakes compound upon themselves. I would say there is a big, fundamental shift between when I was trying to learn what other people had done before me, to when I started trying out newer ideas on my own. 

More importantly, when I started playing, the rolling technique didn’t even exist, so there were completely different ideas about the limits of the game. When we started learning rolling, we were all on our own. Each of us sort of had to try things by ourselves and figure out this new black box. There’s probably an analogy here to some major scientific discovery. 

Q: Spatial awareness and geometric reasoning are a pretty big part of the game — are there ways Tetris has spilled over into other areas of your life? Are you really good at packing for trips?



A: The skills that I have learned from Tetris are pretty hyper-specialized, so no, I am not any better at packing. I will say, in somewhat vague terms, that my involvement with classic Tetris has gotten me an internship; as I wanted to learn more about it, that extends to how the game is programmed. I had to learn assembly and then ROM hacking — which I used in projects that I talked about during interviews.


Q: How does one actually train for a classic Tetris competition?



A: The best way to practice is to just play the game a lot! There is a reinforcement learning analogy: you’re employing a lot of brute force, but always analyzing yourself and your areas of weakness. 


Q: One of the most touching moments of the video was when, immediately after winning, you embraced the runner-up — you both looked so genuinely happy for each other. Have you developed many relationships through Tetris, and is it common for competitors to become friends? 



A: In the really huge competitive video gaming tournaments, with hundreds-of-thousands or million-dollar prize pools, friendships are less common, because those are essentially businesses competing against each other. Classic Tetris is a small community; I know just about everyone who competed in this last tournament. Sidnev, who I played in this final — we’ve played together many times, and this is the culmination of all that. 


Q: Now that you’re the world champion, what will you do to celebrate?

A: Nothing really feels different! I may use this as closure for my experience with the competitive side of the game. I’ve been liking the idea of moving towards the behind-the-scenes infrastructure, like running classic Tetris websites, or running tournaments. 


Q: Does that mean you won’t be defending your title next year?


A: Oh, I absolutely will! There’s still a lot left in the game that I want to accomplish. One achievement I’m chasing is the “game crash.” It’s a very difficult goal. You have to play the game for longer than anyone has ever played before, eventually getting to a point where the loops that calculate your scoring take too long. We’ve known about the possibility of game crash for around a decade, but no human has ever gotten there — the current record is around 1,400 lines, and you reach the crash at 1,490 lines. Arguably, I would be the person most likely to get to game crash because almost two years ago, I became the first person to achieve the glitched colors levels.

Q: How do you feel the process of working to become the best in the world at something has changed you?



A: Playing Tetris has helped me understand myself more. A lot of people who play tend to struggle with keeping their mental health consistent because it’s a really hard, punishing game. I’ve been lucky to avoid this, and I believe it’s because I’m able to focus not on the big picture, but instead on the little things — for instance, making this next placement as perfect as possible. I’m able to forget about the whole competition, and focus on playing my favorite game in front of a bunch of people, and showing how cool it is. 

I think a lot about something said by Jonas Neubauer, who was arguably one of the greatest Tetris players of all time. It’s engraved on the CTWC trophy: “If you're a high visibility player, it's on you to move the community in a positive direction.”

de MIT News https://ift.tt/PcItWxF

Turning engineers into well-rounded communicators

For MIT engineering students and postdocs, tasks like writing grant proposals, applying to jobs, and presenting research findings require not only technical expertise but also the ability to clearly communicate.

For the last 10 years, the MIT School of Engineering’s Communication Lab has helped students and postdocs develop those communication skills. Using principles like understanding your audience, identifying your purpose, and presenting a single, overarching theme, the lab’s students have advanced their academic and professional careers.

“Learning communications ideas in the abstract can be difficult for engineering and science students, who are communicating very specific types of information and stories to diverse audiences,” says Communication Lab Director Diana Chien PhD ’16. “What makes it stick for them is learning about communication in the context of a particular discipline — for example, how should I portray data about a specific engineering experiment I ran for an audience of peers in my field?”

The lab, known as the Comm Lab to most students, works through a peer-to-peer coaching model in which student and postdoc Communication Fellows host one-on-one sessions to help their peers through specific communications tasks.

“We really believe in empowering trainees to be part of the educational experience,” Chien says. “We heavily invest in training the graduate students and postdocs who work with us, because the Communication Fellows really drive our impact.”

The Communication Lab currently has 85 fellows that have worked with thousands of undergraduate, graduate, and postdocs since the program’s founding. In addition to training students, the Comm Lab also hosts workshops, publishes online articles, and offers other resources to help trainees with their written, oral, and visual communication.

“I always wanted to be that person that could bridge the gap between deep technical engineering innovation and the commercial side,” Jordan Alford ’20 said at the Comm Lab’s recent 10th anniversary celebration. “The Comm Lab has really helped me become the type of engineer that can communicate what’s going on effectively so that everyone can understand me, no matter their background.”

Meeting students where they are

Elise Strobach SM ’17, PhD ’20 was part of the first cohort of mechanical engineering Communication Fellows when the Lab expanded in 2017.

“The Comm Lab helped me build up skills around talking to diverse audiences about my research,” she says. “One of the big things they teach you right away is as a Communication Fellow is that you’re not there to teach somebody how to do a thing. You’re there to mentor them so they can do it for themselves.”

Often students come to the Communication Lab seeking help for a specific assignment, like giving a presentation at a conference. Coaching sessions will then address the assignment, but in the context of broader communications principles. In this way the Lab aligns its coaching with student deadlines to meet them where they need help most immediately.

“It’s about finding opportunities to inject educational principles into a project that somebody is already working on,” Chien says. “We use the urgency behind students’ needs to provide educational support.”

The Communication Lab’s workshops are organized around specific tasks, like writing research papers and applying to jobs. In one recent workshop, the lab helped trainees apply for faculty positions, which can be a confusing process.

“We try to demystify some of these academic processes that are very closed-door,” Chien says. “We also help people a ton with field-specific applications in general.”

The coaches also benefit from the sessions.

“As a PhD candidate, you’re doing experiments all the time, and it can be difficult to feel like you’re making progress,” says Chien, who also served a fellow while at MIT. “Getting positive feedback can be transformative for coaches to see themselves as having an impact.”

Strobach took the presentations she developed in the Comm Lab and used them as part of a speaker series her lab group hosts. From then on, when certain members of her lab were preparing to present at a conference or other event, they’d meet with Strobach beforehand.

“When I think about my experiences leading teams, a lot of those Comm Lab experiences were building blocks for me,” says Strobach, who founded the company AeroShield while participating in the Comm Lab program. “I don’t think I would’ve had the confidence to do that before Comm Lab.”

Chien says Strobach’s experience is common.

“We find that communication as a skill can be a gateway for students seeking out leadership experiences,” Chien says. “Communication can improve your confidence and put you in leadership positions. Many fellows aren’t necessarily approaching the fellowship as a leadership opportunity, but it ends up becoming one.”

The online materials the Communication Lab puts out have provided another way to support students. One of the Lab’s most popular resources are communications kits (“CommKits”) tailored to different engineering disciplines. To date, the CommKits have been viewed more than a million times by people from all over the world.

Some of those people turned out not to be students but communications professionals at other schools. A few years ago, the lab started getting contacted by people trying to start similar programs at their universities. In response, the Communication Lab Summer Institute was launched, which provides students and professionals from any institution with a four-day workshop to help them launch science communication training initiatives.

Setting students up for success

By connecting students and postdocs for one-on-one sessions, the Comm Lab also gives participants an outlet to seek help.

“The peer-to-peer format encourages experience sharing,” Strobach says. “It makes it less about micromanaging and just telling someone what to do, and more about seeing where someone’s struggling and showing them something about my prior experience where I’ve struggled and how the Comm Lab has taught me to address those struggles.”

Students also said the Comm Lab helped them with feelings of imposter syndrome and provided a welcoming atmosphere for learning new skills.

“There was a sense of psychological safety there,” Strobach says of the Lab. “As I was trained to become a fellow, it felt very natural. I really valued that, especially being a female in mechanical engineering. It made my time at MIT more impactful, and set me up for the next step in my career.”

de MIT News https://ift.tt/BVcreUE

A marvel in masonry shows the art of the possible

In the Hudson River Valley, on a hill inside the Storm King Art Center, a new addition to the country’s leading outdoor sculpture collection was unveiled this fall. “Lookout,” by the eminent American sculptor Martin Puryear, is a beguiling, domed brick structure with confounding curves, a walk-in entrance, and 90 apertures.

The sculpture “could be the most amazing thing Martin’s ever done,” the noted curator John Elderfield told The New York Times.

“Lookout” also raises a question: How do you assemble an entirely curving building using rectilinear bricks, of all materials?

Answer: First, assemble a team of leading masonry experts to work on it — including several individuals who are MIT faculty, alumni, and students.

The vision of “Lookout” comes from Puryear, who is known for creating evocative shapes and deploying striking materials; it was partly inspired by ancient Nubian masonry techniques. But bricks are best suited for building rectangular structures. They work well in compression — stacking on to each other — but can pull apart when stretched into other shapes. Constructing the completely curvilinear “Lookout” in brick was a major challenge.

As it happens, several MIT-connected people played a key role assembling the sculpture, including John Ochsendorf, an MIT professor and expert in the design of unconventional structures; Lara Davis MArch ’10, a practicing architect and masonry specialist, and the project’s lead mason; Rebecca Buntrock MEng ’10, an engineer with the consulting firm Silman; and a group of students Ochsendorf assembled to analyze and help develop the structure. They joined with Puryear’s studio, collaborating closely with his lead studio assistant, Rob Horton, as well as other building specialists, in crafting “Lookout.”

After more than a year of on-site construction, Puryear told MIT News, it “felt like a final miracle” to see the sculpture completed with the aid of the team — whose members in turn were delighted to help make “Lookout” a reality.

“Martin’s not just a great artist, he’s a great builder, he’s a great craftsman, he’s a maker in every sense of the word,” says Ochsendorf. “He also is devoted to materials and to truth and to an exploration of form. So, we explored this together.”

Adds Davis: “The genius of it is the combination of this really elegant apparent simplicity and technical sophistication.”

A common language around the brick

Puryear is one of the most accomplished sculptors in America. Having first gained notice in the 1970s, he now has a long list of major exhibits, installations, and awards to his name. In 2007, the Museum of Modern Art in New York featured a 30-year retrospective of his work; Puryear’s sculptures also represented the U.S. at the 2019 Venice Biennale for the Arts, and his work was featured multiple times in the Whitney Biennial. He has created many works standing as public installations. In 2011, Puryear was awarded the National Medal of Arts. He has also earned a MacArthur Fellowship and a Guggenheim Fellowship.  

Puryear had already been envisioning “Lookout” for many years when a 2019 conversation with Ochsendorf, the MIT Class of 1942 Professor with appointments in the departments of Architecture and Civil and Environmental Engineering, helped hammer out a framework for building the massive sculpture.

“That meeting resulted in a collaborative structural solution that was almost instantaneous,” Puryear says.  

As Puryear and Ochsendorf determined, “Lookout” would have two brick layers that would be largely self-supporting during construction. The masonry would reflect the Nubian practice of placing bricks at angles with rapidly drying mortar to create a unique curving shape. They also decided to place a steel grid in between those layers for establishing the complex geometry and for long-term reinforcement. Still, the crucial factor would be an extremely precise bricklaying process.

Certainly, Puryear was talking to the right person in Ochsendorf, an engineer and designer with a distinctive niche studying indigenous, ancient, and unusual architectural practices. Ochsendorf’s 2010 book “Guastavino Vaulting: The Art of Structural Tile,” examines the distinctive 19th-century masonry featured in New York’s Grand Central Terminal, Ellis Island, the National Museum of Natural History, and other prominent buildings. And Ochsendorf helped provide engineering analysis for the Sean Collier Memorial, designed by J. Meejin Yoon, on the MIT campus.

“I’ve long been interested in Nubian vault construction,” Ochsendorf says. “But it’s not every artist who knows about it. So, there was a common language around the brick and what it can do in three dimensions, and how we can get there.”

Ochsendorf adds: “When Martin feels something can be done, it probably can be done. The joy for me as someone who has studied masonry structures throughout cultures, time, and space, was that when I saw the form, I thought, yes, this can be done, but it’s not going to be easy, and we’re going to need a lot of talent to make this happen. That led to pulling together a team of MIT students and alums.”

Matching the mortar

For starters, Ochsendorf introduced Puryear to Davis, who became the lead mason on the project and produced dozens of detailed drawings of the structure. Davis had been a mason before earning her masters at MIT, where she studied with Ochsendorf. She then worked in India for a decade before returning to the U.S. and settling near Storm King.

“Lara was living, almost miraculously, only 30 minutes from my studio when we met, and she immediately brought her architectural background, her engineering training, and her hands-on bricklaying expertise to the project,” Puryear explains.

Indeed, a crucial part of the construction involved finding the right match of brick and mortar. Nothing with the shape of “Lookout” could use off-the-shelf materials. Instead, Davis carefully formulated a blend of natural cements to properly bind the bricks, testing out each option through a half-scale mockup built with Puryear’s studio.

“The material properties of the mortar and the brick together really had to be perfectly tuned in order for it be stable and cohesive as we were building it,” Davis says. “As we were building it up, there were points where I felt, wow, the shape is so unusual for masonry.”

Davis, in turn, also worked with masons Scott Cafarella and Mario Magana, of Hudson Valley Mason Works, to lay the bricks. The firm KC Fabrications, founded by Christopher Powers and Kurt Wulfmeyer, built the steel framework. Another of Puryear’s studio assistants, Aaron Getman-Pickering, also collaborated on the project.

Well before construction started, the team also performed multiple engineering assessments to see whether “Lookout” was viable — or if it would create too much structural tension for masonry. One of these studies was performed by Buntrock, a senior associate at Silman whose MIT master’s degree was in high-performance structures.

Additional analysis and planning was led by Ochsendorf, and helped by a team of six students: Sabrina Madera ’21, Jaime Osuna ’21, and MArch candidates Katie Rotman, Justin Brazier, Tejumola Bayowa, and Charlie Janson. Together they conducted further modeling and digital planning of the construction, down to the number of bricks the sculpture would need.  

This kind of “early digital modeling” helped provide “a better view of what it might actually look like,” says Madera. “The gratification of seeing it built was just so wonderful.”

As the process of testing mock-ups indicates, the construction process itself was a highly collaborative one between the assembled team and Puryear’s studio, as the group worked to refine its practices and find solutions for any issue that came up. After many months of work taking place on scaffolding, the careful lifting of heavy component pieces into place, and precise measurements of the structure in progress, “Lookout” was completed by the gifted crew of masons and builders.

“It’s almost like a piece that we have spent our whole careers training for,” Davis says.

As “Lookout” reinforces, great sculpture starts with a vision, and gets completed after careful, exacting, hands-on craft work.

“It was a great intellectual challenge for us, but we were also relying on historical precedent,” Ochsendorf says. “Like Martin, we knew it was possible.”

de MIT News https://ift.tt/AI5qmH6

School of Engineering third quarter 2023 awards

Faculty and researchers across MIT’s School of Engineering receive many awards in recognition of their scholarship, service, and overall excellence. The School of Engineering periodically recognizes their achievements by highlighting the honors, prizes, and medals won by faculty and research scientists working in our academic departments, labs, and centers.

• Jacopo Buongiorno in the Department of Nuclear Science and Engineering was selected as a 2023 NURETH Fellow on Aug. 1.
   
• Joel Emer in the Department of Electrical Engineering and Computer Science received the 2023 B. Ramakrishna Rau Award on Sept. 25.
   
• James Fujimoto in the Department of Electrical Engineering and Computer Science received the 2023 Lasker-DeBakey Clinical Medical Research Award on Sept. 26.
   
• Leonid A. Mirny in the Institute for Medical Science and Engineering was named a 2023 Simons Investigator on Aug. 28.
   
• Dava Newman in the Department of Aeronautics and Astronautics received an Honorary Doctorate from the Royal College of Art, London, U.K. on Sept. 13.
   
• Greg Stephanopoulos in the Department of Chemical Engineering received the James E. Bailey Award on Sept. 27.
   
• Gerald Sussman in the Department of Electrical Engineering and Computer Science received the IEEE Major Education Innovation Award on July 19.
   
• Vinod Vaikuntanathan in the Department of Electrical Engineering and Computer Science received the 2023 International Association of Cryptologic Research Test-of-Time Award on Aug. 8.
   
• Bilge Yildiz in the Department of Nuclear Science and Engineering was selected as a 2023 Fellow of the Electrochemical Society on Aug. 1 and was elected to the Austrian Academy of Sciences on May 1.
   
• Sidney Yip in the Department of Nuclear Science and Engineering received the Monie A. Ferst Award on Sept. 29.

de MIT News https://ift.tt/4mn3VBC

Making genetic prediction models more inclusive

While any two human genomes are about 99.9 percent identical, genetic variation in the remaining 0.1 percent plays an important role in shaping human diversity, including a person’s risk for developing certain diseases.

Measuring the cumulative effect of these small genetic differences can provide an estimate of an individual’s genetic risk for a particular disease or their likelihood of having a particular trait. However, the majority of models used to generate these “polygenic scores” are based on studies done in people of European descent, and do not accurately gauge the risk for people of non-European ancestry or people whose genomes contain a mixture of chromosome regions inherited from previously isolated populations, also known as admixed ancestry.

In an effort to make these genetic scores more inclusive, MIT researchers have created a new model that takes into account genetic information from people from a wider diversity of genetic ancestries across the world. Using this model, they showed that they could increase the accuracy of genetics-based predictions for a variety of traits, especially for people from populations that have been traditionally underrepresented in genetic studies.

“For people of African ancestry, our model proved to be about 60 percent more accurate on average,” 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. “For people of admixed genetic backgrounds more broadly, who have been excluded from most previous models, the accuracy of our model increased by an average of about 18 percent.”

The researchers hope their more inclusive modeling approach could help improve health outcomes for a wider range of people and promote health equity by spreading the benefits of genomic sequencing more widely across the globe.

“What we have done is created a method that allows you to be much more accurate for admixed and ancestry-diverse individuals, and ensure the results and the benefits of human genetics research are equally shared by everyone,” says MIT postdoc Yosuke Tanigawa, the lead and co-corresponding author of the paper, which appears today in open-access form in the American Journal of Human Genetics. The researchers have made all of their data publicly available for the broader scientific community to use.

More inclusive models

The work builds on the Human Genome Project, which mapped all of the genes found in the human genome, and on subsequent large-scale, cohort-based studies of how genetic variants in the human genome are linked to disease risk and other differences between individuals.

These studies showed that the effect of any individual genetic variant on its own is typically very small. Together, these small effects add up and influence the risk of developing heart disease or diabetes, having a stroke, or being diagnosed with psychiatric disorders such as schizophrenia.

“We have hundreds of thousands of genetic variants that are associated with complex traits, each of which is individually playing a weak effect, but together they are beginning to be predictive for disease predispositions,” Kellis says.

However, most of these genome-wide association studies included few people of non-European descent, so polygenic risk models based on them translate poorly to non-European populations. People from different geographic areas can have different patterns of genetic variation, shaped by stochastic drift, population history, and environmental factors — for example, in people of African descent, genetic variants that protect against malaria are more common than in other populations. Those variants also affect other traits involving the immune system, such as counts of neutrophils, a type of immune cell. That variation would not be well-captured in a model based on genetic analysis of people of European ancestry alone.

“If you are an individual of African descent, of Latin American descent, of Asian descent, then you are currently being left out by the system,” Kellis says. “This inequity in the utilization of genetic information for predicting risk of patients can cause unnecessary burden, unnecessary deaths, and unnecessary lack of prevention, and that's where our work comes in.”

Some researchers have begun trying to address these disparities by creating distinct models for people of European descent, of African descent, or of Asian descent. These emerging approaches assign individuals to distinct genetic ancestry groups, aggregate the data to create an association summary, and make genetic prediction models. However, these approaches still don’t represent people of admixed genetic backgrounds well.

“Our approach builds on the previous work without requiring researchers to assign individuals or local genomic segments of individuals to predefined distinct genetic ancestry groups,” Tanigawa says. “Instead, we develop a single model for everybody by directly working on individuals across the continuum of their genetic ancestries.”

In creating their new model, the MIT team used computational and statistical techniques that enabled them to study each individual’s unique genetic profile instead of grouping individuals by population. This methodological advancement allowed the researchers to include people of admixed ancestry, who made up nearly 10 percent of the UK Biobank dataset used for this study and currently account for about one in seven newborns in the United States.

“Because we work at the individual level, there is no need for computing summary-level data for different populations,” Kellis says. “Thus, we did not need to exclude individuals of admixed ancestry, increasing our power by including more individuals and representing contributions from all populations in our combined model.”

Better predictions

To create their new model, the researchers used genetic data from more than 280,000 people, which was collected by UK Biobank, a large-scale biomedical database and research resource containing de-identified genetic, lifestyle, and health information from half a million U.K. participants. Using another set of about 81,000 held-out individuals from the UK Biobank, the researchers evaluated their model across 60 traits, which included traits related to body size and shape, such as height and body mass index, as well as blood traits such as white blood cell count and red blood cell count, which also have a genetic basis.

The researchers found that, compared to models trained only on European-ancestry individuals, their model’s predictions are more accurate for all genetic ancestry groups. The most notable gain was for people of African ancestry, who showed 61 percent average improvements, even though they only made up about 1.5 percent of samples in UK Biobank. The researchers also saw improvements of 11 percent for people of South Asian descent and 5 percent for white British people. Predictions for people of admixed ancestry improved by about 18 percent.

“When you bring all the individuals together in the training set, everybody contributes to the training of the polygenic score modeling on equal footing,” Tanigawa says. “Combined with increasingly more inclusive data collection efforts, our method can help leverage these efforts to improve predictive accuracy for all.”

The MIT team hopes its approach can eventually be incorporated into tests of an individual’s risk of a variety of diseases. Such tests could be combined with conventional risk factors and used to help doctors diagnose disease or to help people manage their risk for certain diseases before they develop.

“Our work highlights the power of diversity, equity, and inclusion efforts in the context of genomics research,” Tanigawa says.

The researchers now hope to add even more data to their model, including data from the United States, and to apply it to additional traits that they didn’t analyze in this study.

“This is just the start,” Kellis says. “We can’t wait to see more people join our effort to propel inclusive human genetics research.”

The research was funded by the National Institutes of Health.

de MIT News https://ift.tt/2Vwt4nU

How adults understand what kids are saying

When babies first begin to talk, their vocabulary is very limited. Often one of the first sounds they generate is “da,” which may refer to dad, a dog, a dot, or nothing at all.

How does an adult listener make sense of this limited verbal repertoire? A new study from MIT and Harvard University researchers has found that adults’ understanding of conversational context and knowledge of mispronunciations that children commonly make are critical to the ability to understand children’s early linguistic efforts.

Using thousands of hours of transcribed audio recordings of children and adults interacting, the research team created computational models that let them start to reverse engineer how adults interpret what small children are saying. Models based on only the actual sounds children produced in their speech did a relatively poor job predicting what adults thought children said. The most successful models made their predictions based on large swaths of preceding conversations that provided context for what the children were saying. The models also performed better when they were retrained on large datasets of adults and children interacting.

The findings suggest that adults are highly skilled at making these context-based interpretations, which may provide crucial feedback that helps babies acquire language, the researchers say.

“An adult with lots of listening experience is bringing to bear extremely sophisticated mechanisms of language understanding, and that is clearly what underlies the ability to understand what young children say,” says Roger Levy, a professor of brain and cognitive sciences at MIT. “At this point, we don’t have direct evidence that those mechanisms are directly facilitating the bootstrapping of language acquisition in young children, but I think it’s plausible to hypothesize that they are making the bootstrapping more effective and smoothing the path to successful language acquisition by children.”

Levy and Elika Bergelson, an associate professor of psychology at Harvard, are the senior authors of the study, which appears today in Nature Human Behavior. MIT postdoc Stephan Meylan is the lead author of the paper.

Adult listening skills are critical

While many studies have investigated how children learn to speak, in this project, the researchers wanted to flip the question and study how adults interpret what children say.

“While people have looked historically at a number of features of the learner, and what is it about the child that allows them to learn things from the world, very little has been done to look at how they are understood and how that might influence the process of language acquisition,” Meylan says.

Previous research has shown that when adults speak to each other, they use their beliefs about how other people are likely to talk, and what they’re likely to talk about, to help them understand what their conversational partner is saying. This strategy, known as “noisy channel listening,” makes it easier for adults to handle the complex task of deciphering the acoustic sounds they’re hearing, especially in environments where voices are muffled or there is a lot of background noise, or when speakers have different accents.

In this study, the researchers explored whether adults can also apply this technique to parsing the often seemingly nonsensical utterances produced by children who are learning to talk.

“This problem of interpreting what we hear is even harder for child language than ordinary adult language understanding, which is actually not that easy either, even though we’re very good at it,” Levy says.

For this study, the researchers made use of datasets originally generated at Brown University in the early 2000s, which contain hundreds of hours of transcribed conversations between children ages 1 to 3 and their caregivers. The data include both phonetic transcriptions of the sounds produced by the children and the text of what the transcriber believed the child was trying to say.

The researchers used other datasets of child language (which included about 18 million spoken words) to train computational language models to predict what words the children were saying in the original dataset, based on the phonetic transcription. Using neural networks, they created many different models, which varied in the sophistication of their knowledge of conversational topics, grammar, and children’s mispronunciations. They also manipulated how much of the conversational context each model was allowed to analyze before making its predictions of what the children said. Some models took into account just one or two words spoken before the target word, while others were allowed to analyze up to 20 previous utterances in the exchange.

The researchers found that using the acoustics of what the child said alone did not lead to models that were particularly accurate at predicting what adults thought children said. The models that did best used very rich representations of conversational topics, grammar, and beliefs about what words children are likely to say (ball, dog or baby, rather than mortgage, for example). And much like humans, the models’ predictions improved as they were allowed to consider larger chunks of previous exchanges for context.

A feedback system

The findings suggest that when listening to children, adults base their interpretation of what a child is saying on previous exchanges that they have had. For example, if a dog had been mentioned earlier in the conversation, “da” was more likely to be interpreted by an adult listener as “dog.”

This is an example of a strategy that humans often use in listening to other adults, which is to base their interpretation on “priors,” or expectations based on prior experience. The findings also suggest that when listening to children, adult listeners incorporate expectations of how children commonly mispronounce words, such as “weed” for “read.”

The researchers now plan to explore how adults’ listening skills, and their subsequent responses to children, may help to facilitate children’s ability to learn language.

“Most people prefer to talk to others, and I think babies are no exception to this, especially if there are things that they might want, either in a tangible way, like milk or to be picked up, but also in an intangible way in terms of just the spotlight of social attention,” Bergelson says. “It’s a feedback system that might push the kid, with their burgeoning social skills and cognitive skills and everything else, to continue down this path of trying to interact and communicate.”

One way the researchers hope to study this interplay between child and adult is by combining computational models of how children learn language with the new model of how adults respond to what children say.

“We now have this model of an adult listener that we can plug into models of child learners, and then those learners can leverage the feedback provided by the adult model,” Meylan says. “The next frontier is trying to understand how kids are taking the feedback that they get from these adults and build a model of what these children expect that an adult would understand.”

The research was funded by the National Science Foundation, the National Institutes of Health, and a CONVO grant to MIT’s Department of Brain and Cognitive Sciences from the Simons Center for the Social Brain.

de MIT News https://ift.tt/i1Iz4A5

Shape-shifting fiber can produce morphing fabrics

Instead of needing a coat for each season, imagine having a jacket that would dynamically change shape so it becomes more insulating to keep you warm as the temperature drops.

A programmable, actuating fiber developed by an interdisciplinary team of MIT researchers could someday make this vision a reality. Known as FibeRobo, the fiber contracts in response to an increase in temperature, then self-reverses when the temperature decreases, without any embedded sensors or other hard components.

The low-cost fiber is fully compatible with textile manufacturing techniques, including weaving looms, embroidery, and industrial knitting machines, and can be produced continuously by the kilometer. This could enable designers to easily incorporate actuation and sensing capabilities into a wide range of fabrics for myriad applications, such as programmable compression garments that could aid in post-surgery recovery.

The fibers can also be combined with conductive thread, which acts as a heating element when electric current runs through it. In this way, the fibers actuate using electricity, which offers a user digital control over a textile’s form. For instance, a fabric could change shape based on any piece of digital information, such as readings from a heart rate sensor.

“We use textiles for everything. We make planes with fiber-reinforced composites, we cover the International Space Station with a radiation-shielding fabric, we use them for personal expression and performance wear. So much of our environment is adaptive and responsive, but the one thing that needs to be the most adaptive and responsive — textiles — is completely inert,” says Jack Forman, a graduate student in the Tangible Media Group and the Center for Bits and Atoms in the MIT Media Lab, and lead author of a paper on the actuating fiber.

He is joined on the paper by 11 other researchers at MIT and Northeastern University, including his advisors, Professor Neil Gershenfeld, who leads the Center for Bits and Atoms, and Hiroshi Ishii, the Jerome B. Wiesner Professor of Media Arts and Sciences and director of the Tangible Media Group. The research will be presented at the ACM Symposium on User Interface Software and Technology.

Morphing materials

Current shape-changing fibers have pitfalls that have largely prevented them from being incorporated into textiles beyond laboratory settings.

One fiber, known as a shape-changing alloy, only contracts by about 5 percent, doesn’t self-reverse, and often stops working after a handful of actuations. Another, called a McKibben actuator, is pneumatically driven and requires an air compressor to actuate.

The MIT researchers wanted a fiber that could actuate silently and change its shape dramatically, while being compatible with common textile manufacturing procedures. To achieve this, they used a material known as liquid crystal elastomer (LCE).

A liquid crystal is a series of molecules that can flow like liquid, but when they’re allowed to settle, they stack into a periodic crystal arrangement. The researchers incorporate these crystal structures into an elastomer network, which is stretchy like a rubber band.

As the LCE material heats up, the crystal molecules fall out of alignment and pull the elastomer network together, causing the fiber to contract. When the heat is removed, the molecules return to their original alignment, and the material to its original length, Forman explains.

By carefully mixing chemicals to synthesize the LCE, the researchers can control the final properties of the fiber, such as its thickness or the temperature at which it actuates.

They perfected a preparation technique that creates LCE fiber which can actuate at skin-safe temperatures, making it suitable for wearable fabrics. Researchers had been unable to accomplish this with other LCE fibers, Forman says.

“There are a lot of knobs we can turn. It was a lot of work to come up with this process from scratch, but ultimately it gives us a lot of freedom for the resulting fiber,” he adds.

However, the researchers discovered that making fiber from LCE resin is a finicky process. Existing techniques often result in a fused mass that is impossible to unspool.

Researchers are also exploring other ways to make functional fibers, such as by incorporating hundreds of microscale digital chips into a polymer, utilizing an activated fluidic system, or including piezoelectric material that can convert sound vibrations into electrical signals.

Fiber fabrication

Forman built a machine using 3D-printed and laser-cut parts and basic electronics to overcome the fabrication challenges. He initially built the machine as part of the graduate-level course MAS.865 (Rapid-Prototyping of Rapid-Prototyping Machines: How to Make Something that Makes [almost] Anything).

To begin, the thick and viscous LCE resin is heated, and then slowly squeezed through a nozzle like that of a glue gun. As the resin comes out, it is cured carefully using UV lights that shine on both sides of the slowly extruding fiber.

If the light is too dim, the material will separate and drip out of the machine, but if it is too bright, clumps can form, which yields bumpy fibers.

Then the fiber is dipped in oil to give it a slippery coating and cured again, this time with UV lights turned up to full blast, creating a strong and smooth fiber. Finally, it is collected into a top spool and dipped in powder so it will slide easily into machines for textile manufacturing.

From chemical synthesis to finished spool, the process takes about a day and produces approximately a kilometer of ready-to-use fiber.

“At the end of the day, you don’t want a diva fiber. You want a fiber that, when you are working with it, falls into the ensemble of materials — one that you can work with just like any other fiber material, but then it has a lot of exciting new capabilities,” Forman says.

Creating such a fiber took a great deal of trial and error, as well as the collaboration of researchers with expertise in many disciplines, from chemistry to mechanical engineering to electronics to design.

The resulting fiber, called FibeRobo, can contract up to 40 percent without bending, actuate at skin-safe temperatures, and be produced with a low-cost setup for 20 cents per meter, which is about 60 times cheaper than commercially available shape-changing fibers.

The fiber can be incorporated into industrial sewing and knitting machines, as well as nonindustrial processes like hand looms or manual crocheting, without the need for any process modifications.

The MIT researchers used FibeRobo to demonstrate several applications, including an adaptive sports bra made by embroidery that tightens when the user begins exercising.

They also used an industrial knitting machine to create a compression jacket for Forman’s dog, whose name is Professor. The jacket would actuate and “hug” the dog based on a Bluetooth signal from Forman’s smartphone. Compression jackets are commonly used to alleviate the separation anxiety a dog can feel while its owner is away.

In the future, the researchers want to adjust the fiber’s chemical components so it can be recyclable or biodegradable. They also want to streamline the polymer synthesis process so users without wet lab expertise could make it on their own.

Forman is excited to see the FibeRobo applications other research groups identify as they build on these early results. In the long run, he hopes FibeRobo can become something a maker could buy in a craft store, just like a ball of yarn, and use to easily produce morphing fabrics.

“LCE fibers come to life when integrated into functional textiles. It is particularly fascinating to observe how the authors have explored creative textile designs using a variety of weaving and knitting patterns,” says Lining Yao, the Cooper-Siegel Associate Professor of Human Computer Interaction at Carnegie Mellon University, who was not involved with this work.

This research was supported, in part, by the William Asbjornsen Albert Memorial Fellowship, the Dr. Martin Luther King Jr. Visiting Professor Program, Toppan Printing Co., Honda Research, Chinese Scholarship Council, and Shima Seiki. The team included Ozgun Kilic Afsar, Sarah Nicita, Rosalie (Hsin-Ju) Lin, Liu Yang, Akshay Kothakonda, Zachary Gordon, and Cedric Honnet at MIT; and Megan Hofmann and Kristen Dorsey at Northeastern University.

de MIT News https://ift.tt/JhWLAV0

Morris Chang ’52, SM ’53 describes the secrets of semiconductor success

Groundbreaking technologist Morris Chang ’52, SM ’53 discussed the key elements behind Taiwan’s long-term ascendancy in semiconductor manufacturing, while speaking to a large campus audience in an MIT talk on Tuesday.

Chang is the influential founder and former longtime head of TSMC, the Taiwan Semiconductor Manufacturing Company, which has become the world’s leading microchip maker. Chang started the firm in 1987, and since then it has helped reshape the industry by making Taiwan a crucial center of production and by focusing on manufacturing while chip design occurs elsewhere.

In his remarks, Chang, whose career spans the history of the semiconductor industry, gave a broad overview of the development of chip manufacturing, then focused on some of the factors that have helped TSMC and Taiwan thrive. In his view, this includes a healthy supply of talent, in the form of engineers and other technical employees willing to work in manufacturing; low turnover of employees; a geographical concentration of industry manufacturing in Taiwan; and the “experience curve theory,” in which accumulated manufacturing experience leads to lower production costs.

“Why is TSMC successful in Taiwan?” asked Chang. “Because TSMC also gets good, well-trained technicians, and even well-trained operators from a lot of trade schools in Taiwan. … Their students aspire to make a good living as technicians.”

Chang also argued that the process of learning by doing, in which manufacturers can reduce costs by improving their processes, is predicated on having production centered in one place, with connectivity among workers in common conditions.

“It works, the learning curve, the experience curve, it works only when you have a common location,” Chang said. “Learning is local.”

More broadly, Chang noted, prominence in semiconductor manufacturing “seems to be related to the status of economic development of that country. Frankly, the advantages that Taiwan enjoys today … were enjoyed by the U.S. in the ’50s and ’60s.” Decades in the future, he suggested, there could be a rise in semiconductor activity in India, Vietnam, or Indonesia, depending on the way circumstances evolve.

Chang’s talk, “Lessons of a Life in Semiconductor Manufacturing, from Texas to Taiwan,” was delivered to a capacity audience of more than 425 people in MIT’s Room 10-250. The event was part of the Manufacturing@MIT Distinguished Speaker Series.

Chang was born in China in 1931, left the country in the late 1940s, and earned his BS and MS in mechanical engineering from MIT. He later received a PhD in Electrical Engineering from Stanford University. Chang entered the semiconductor business by taking a job at Sylvania in the mid-1950s. He moved to Texas Instruments — then a chip-making power — in 1958, rising to industry prominence during a quarter-century tenure there. Chang also became a U.S. citizen in 1962. In the 1980s, he was invited to work on the development of industrial and technology in Taiwan, and soon thereafter launched TSMC. Over time, TSMC has become a juggernaut, sustaining its success and growing to become the world leader in the field.

At Tuesday’s event, Chang was introduced by MIT Provost Cynthia Barnhart, who emphasized the enduring connections he has built at the Institute. Chang is a life member emeritus of the MIT Corporation and has been an important supporter of the Institute. The renovated Building E52, which houses the Department of Economics, the Samberg Conference Center, and offices of the MIT Sloan School of Management, is now the Morris and Sophie Chang Building.

“For MIT, Morris is an extraordinary example of the lasting impact our alumni have on the legacy of innovation at the Institute,” Barnhart said.

At the end of his talk, Chang also briefly discussed geopolitics and chip production. With U.S.-China relations relatively strained, chip manufacturing is one of the areas where they are competing in industrial terms. Meanwhile, the precise nature of China’s policy intentions vis-à-vis Taiwan remain uncertain.

“Without national security, we will lose everything, everything that we value,” Chang said. At the same time, he noted, “By all means, let’s avoid even a Cold War, if we can.”

The event was hosted by the Manufacturing@MIT Working Group, in collaboration with the Department of Mechanical Engineering, the Department of Political Science, Leaders for Global Operations, Microsystems Technology Laboratories, the Industrial Performance Center, MIT.nano, Machine Intelligence for Manufacturing and Operations, the Laboratory for Manufacturing and Productivity, and Mission Innovation X.

de MIT News https://ift.tt/vqaSMwC