A foolproof way to shrink deep learning models

As more artificial intelligence applications move to smartphones, deep learning models are getting smaller to allow apps to run faster and save battery power. Now, MIT researchers have a new and better way to compress models.  It’s so simple that they unveiled it in a tweet last month: Train the model, prune its weakest connections, retrain the model at its fast, early training rate, and repeat, until the model is as tiny as you want.  “That’s it,” says Alex Renda, a PhD student at MIT. “The standard things people do to prune their models are crazy complicated.”  Renda discussed the technique when the International Conference of Learning Representations (ICLR) convened remotely this month. Renda is a co-author of the work with Jonathan Frankle, a fellow PhD student in MIT’s Department of Electrical Engineering and Computer Science (EECS), and Michael Carbin, an assistant professor of electrical engineering and computer science — all members of the Computer Science and Continue reading A foolproof way to shrink deep learning models

Researchers help give robotic arms a steady hand for surgeries

Steady hands and uninterrupted, sharp vision are critical when performing surgery on delicate structures like the brain or hair-thin blood vessels. While surgical cameras have improved what surgeons see during operative procedures, the ‘steady hand’ remains to be enhanced — new surgical technologies, including sophisticated surgeon-guided robotic hands, cannot prevent accidental injuries when operating close to fragile tissue. More details

Robotic Arm Wields UV Light Wand To Disinfect Public Spaces

THE INSTITUTE Properly disinfecting public spaces can help stop the spread of coronavirus, but cleaning crews are often not properly trained how to do so. Also, if the workers don’t wear personal protective equipment, they are at risk for infection. IEEE Fellow Satyandra K. Gupta is leading a research team at the University of Southern California’s Viterbi Center for Advanced Manufacturing in Los Angeles that is building a robotic arm that uses a UV light sanitizer to clean contaminated areas. Gupta is a mechanical engineering professor at the university. The Institute asked him about the project. This interview has been edited and condensed for clarity. What problem are you trying to solve? In order to thoroughly clean cluttered work spaces, janitorial staff must move items so that all high-touch surfaces can be disinfected. This requires a significant amount of time and effort. It also puts the lives of the workers at Continue reading Robotic Arm Wields UV Light Wand To Disinfect Public Spaces

Spinal cord gives bio-bots walking rhythm

Miniature biological robots are making greater strides than ever, thanks to the spinal cord directing their steps. Researchers developed the tiny walking ‘spinobots,’ powered by rat muscle and spinal cord tissue on a soft, 3D-printed hydrogel skeleton. While previous generations of biological robots, or bio-bots, could move forward by simple muscle contraction, the integration of the spinal cord gives them a more natural walking rhythm. More details

Meet Moxie, a Social Robot That Helps Kids With Social-Emotional Learning

The first generation of social home robots (like the first generation of many other new applications of technology) was not particularly successful. A series of high valuations followed by mediocre sales and reviews leading to several company shutdowns has made it much more challenging to develop in this space. And that’s not necessarily a bad thing, if we’re honest—it’s become clear that social home robots are difficult to get right, and any company that wants to make one at this point needs to have both the technical capability and (more importantly) a long-term business case that’s more comprehensive than just making a robot into a “part of the family” or a “best friend.” Today, a social robotics startup called Embodied is launching a new robot called Moxie (no relation to this or this), a social companion for children aged 6ish to 9ish. According to Embodied, Moxie is “designed to help promote social, Continue reading Meet Moxie, a Social Robot That Helps Kids With Social-Emotional Learning

Automating the search for entirely new “curiosity” algorithms

Driven by an innate curiosity, children pick up new skills as they explore the world and learn from their experience. Computers, by contrast, often get stuck when thrown into new environments. To get around this, engineers have tried encoding simple forms of curiosity into their algorithms with the hope that an agent pushed to explore will learn about its environment more effectively. An agent with a child’s curiosity might go from learning to pick up, manipulate, and throw objects to understanding the pull of gravity, a realization that could dramatically accelerate its ability to learn many other things.  Engineers have discovered many ways of encoding curious exploration into machine learning algorithms. A research team at MIT wondered if a computer could do better, based on a long history of enlisting computers in the search for new algorithms.  In recent years, the design of deep neural networks, algorithms that search for solutions Continue reading Automating the search for entirely new “curiosity” algorithms

Akara Robotics Turns TurtleBot Into Autonomous UV Disinfecting Robot

UV disinfection is one of the few areas where autonomous robots can be immediately and uniquely helpful during the COVID pandemic. Unfortunately, there aren’t enough of these robots to fulfill demand right now, and although companies are working hard to build them, it takes a substantial amount of time to develop the hardware, software, operational knowledge, and integration experience required to make a robotic disinfection system work in a hospital.  Conor McGinn, an assistant professor of mechanical engineering at Trinity College in Dublin and co-leader of the Robotics and Innovation Lab (RAIL), has pulled together a small team of hardware and software engineers who’ve managed to get a UV disinfection robot into hospital testing within a matter of just a few weeks. They made it happen in such a short amount of time by building on previous research, collaborating with hospitals directly, and leveraging a development platform: the TurtleBot 2.

3 Questions: Tom Leighton on the major surge in internet traffic triggered by physical distancing

With various physical distancing guidelines in place throughout the world as a means to curb the spread of Covid-19, the internet has experienced a dramatic spike in overall traffic. MIT Professor Tom Leighton is chief executive officer and co-founder of Akamai Technologies, a global content delivery network, cybersecurity, and cloud service company that provides web and internet security services. At MIT he specializes in applied mathematics in the Department of Mathematics and is a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL). The Department of Mathematics Communications spoke to Leighton about his company’s response to the world’s increased reliance on the internet during the Covid-19 pandemic. Q: How is the pandemic changing the way people use the internet? A: The internet has become our lifeline as we face the challenges of working remotely, distance learning, and sheltering in place. Everything has moved online: religious services, movie premieres, commerce Continue reading 3 Questions: Tom Leighton on the major surge in internet traffic triggered by physical distancing