MIT engineers have developed a new training method to help ensure the safe operation of multiagent systems, including robots, search-and-rescue drones and self-driving cars, reports Jijo Malayil for Interesting Engineering. The new approach “doesn’t focus on rigid paths but rather enables agents to continuously map their safety margins—the boundaries within which they must stay,” writes Malayil.
Graduate students Suhan Kim and Yi-Hsuan (Nemo) Hsiao speak with Tech Briefs reporter Andrew Corselli about their work developing insect-sized robots capable of artificial pollination. “Typical drones use electromagnetic motors plus propellers. But, our system is a little different in that we are primarily using an artificial muscle,” explains Kim.
Researchers from MIT and elsewhere have develop insect-sized robots that could one day be used to help with farming practices like artificial pollination, reports Alice Rizzo for Reuters. "These type of robots will open up a very new type of use case," says graduate student Suhan Kim. "We can start thinking of using our robot, if it works well, for tools like indoor farming."
In a letter to the editor of The Boston Globe, Vice President for Research Ian Waitz addresses the importance of research staff at the Institute, noting that “research universities educate through research.” Waitz emphasizes: “At MIT, there has been double-digit real growth in our on-campus research enterprise over the past 11 years along with growth in our graduate student body. With that come more people, and while these staff may not be directly involved in student classroom instruction, the research they conduct is crucial to the hands-on education that MIT students receive and to the real-world solutions that originate at the school.”
Researchers at MIT have developed an insect-like, flying robot capable of performing acrobatic maneuvers and hovering in the air for up to 15 minutes without failing, reports Alex Wilkins for New Scientist. “By having a hugely increased [flying] lifetime, we were able to work on the controller parts so that the robot can achieve precise trajectory tracking, plus aggressive maneuvers like somersaults,” says graduate student Suhan Kim.
MIT engineers have developed “two new control techniques that have enabled them to achieve a world-record single-qubit fidelity of 99.998 percent using a superconducting qubit called fluxonium,” reports Aman Tripathi for Interesting Engineering. “This breakthrough marks a significant step towards the realization of practical quantum computing,” Tripathi notes.
Ars Technica reporter Jacek Krywko spotlights how MIT researchers have developed a new photonic chip that that can “compute the entire deep neural net, including both linear and non-linear operations, using photons.” Visiting scientist Saumil Bandyopadhyay '17, MEng '18, PhD '23 explains that: “We’re focused on a very specific metric here, which is latency. We aim for applications where what matters the most is how fast you can produce a solution. That’s why we are interested in systems where we’re able to do all the computations optically.”
Writing for Wired, Prof. Daniela Rus, director of CSAIL, highlights the future of “physical intelligence, a new form of intelligent machine that can understand dynamic environments, cope with unpredictability, and make decisions in real time.” Rus writes: “Unlike the models used by standard AI, physical intelligence is rooted in physics; in understanding the fundamental principles of the real world, such as cause-and-effect.”
Liquid AI, an MIT startup, is developing technology that “holds the same promise of writing, analyzing, and creating content as its rivals while using far less computing power,” reports Aaron Pressman for The Boston Globe.
Prof. Daron Acemoglu, one of the recipients of the 2024 Nobel Prize in economics, speaks with NPR Planet Money hosts Jeff Guo and Greg Rosalsky about the academic inspirations that led to his award-winning research studying the role of institutions in shaping economies. “In 1980, as I was in middle school, just the beginning of my seventh grade, Turkey suffered a big military coup,” explains Acemoglu. “There were soldiers everywhere, including in our school. Turkey was definitely not a democratic country at the time, and it was also suffering via a series of economic problems. I got interested in exactly these sets of issues.”
Prof. Daniela Rus, director of CSAIL, speaks with NBC Boston reporter Colton Bradford about her work developing a new AI system aimed at making grocery shopping easier, more personalized and more efficient. “I think there is an important synergy between what people can do and what machines can do,” says Rus. “You can think of it as machines have speed, but people have wisdom. Machines can lift heavy things, but people can reason about what to do with those heavy things.”
Using a new technique developed to examine the risks of multimodal large language models used in robots, MIT researchers were able to have a “simulated robot arm do unsafe things like knocking items off a table or throwing them by describing actions in ways that the LLM did not recognize as harmful and reject,” writes Will Knight for Wired. “With LLMs a few wrong words don’t matter as much,” explains Prof. Pulkit Agrawal. “In robotics a few wrong actions can compound and result in task failure more easily.”
Grist reporter Matt Simon spotlights a new study co-authored by MIT researchers that finds “large-scale deployment of long-duration energy storage isn’t just feasible but essential for renewables to reach their full potential, and would even cut utility bills.” Graduate student Martin Staadecker explains: “Battery storage on its own — or what people call short-duration energy storage — is very important. But you can’t just rely on lithium-ion batteries, because it would be very expensive to have enough to actually provide power for an entire week.”
Prof. Armando Solar-Lezama speaks with New York Times reporter Sarah Kessler about the future of coding jobs, noting that AI systems still lack many essential skills. “When you’re talking about more foundational skills, knowing how to reason about a piece of code, knowing how to track down a bug across a large system, those are things that the current models really don’t know how to do,” says Solar-Lezama.
Prof. Daniela Rus, director of CSAIL, and Prof. Russ Tedrake speak with the Financial Times about how advances in AI have made it possible for robots to learn new skills and perform complex tasks. “All these cool things that we only dreamed of, we can now begin to realize,” says Rus. “Now we have to make sure that what we do with all these superpowers is good.”