Alexander Sludds, a graduate student in MIT’s Research Lab for Electronics, joins Megan Cantwell on the Science magazine podcast to discuss his team’s new method for processing data on edge devices, which are devices that connect two networks together.
MIT’s mini cheetah robot was taught how to goal keep using simulation, reports Emmett Smith for Mashable. “The robot was able to block 87.5 percent of the shots taken, which is just slightly above the best professional goalies in the English Premiere League,” writes Smith.
In a new paper, MIT researchers detail the use of reinforcement learning to teach MIT’s mini cheetah robot to play goalie in a soccer match, reports Brian Heater for TechCrunch. “In this work, we focused solely on the goalkeeping task, but the proposed framework can be extended to other scenarios, such as multi-skill soccer ball kicking,” the researchers explain.
Institute Prof. Barbara Liskov discusses the importance of including ethics and foresight as a key parts of computer science education, reports Susan D'Agostino for Inside Higher Ed. “The days of being naïvely technical, which we were for many years, are over,” says Liskov. “We need to open students’ minds so they think about the harm that can come from what they’re doing and so they ask, ‘What could I add that could act as a safeguard?’ It’s more than ethics. They need to think from a different perspective.”
At MIT’s AI Policy Forum Summit, which was focused on exploring the challenges facing the implementation of AI technologies across a variety of sectors, SEC Chair Gary Gensler and MIT Schwarzman College of Computing Dean Daniel Huttenlocher discussed the impact of AI on the world of finance. “If someone is relying on open-AI, that's a concentrated risk and a lot of fintech companies can build on top of it,” Gensler said. “Then you have a node that's every bit as systemically relevant as maybe a stock exchange."
MIT researchers have developed a new in-home device that can help monitor Parkinson’s patients by tracking their gait, reports Hiawatha Bray for The Boston Globe. “We know very little about the brain and its diseases,” says Professor Dina Katabi. “My goal is to develop non-invasive tools that provide new insights about the functioning of the brain and its diseases.”
Popular Science reporter Philip Kiefer writes that MIT researchers have developed an in-home device that could be used to track the progression of symptoms in Parkinson’s patients. “We can’t really ask patients to come to the clinic every day or every week,” explains graduate student Yingcheng Liu. “This technology gives us the possibility to continuously monitor patients, and provide more objective assessments.”
TechCrunch reporter Kyle Wiggers spotlights DynamoFl, a company founded by Christian Lau PhD ’20 and Vaikkunth Mugunthan PhD ’22 that is developing a federated learning platform, a technique for preserving data privacy in AI systems.
Researchers at MIT have found that “Salet” is the best starting word for Wordle, reports Erik Kain for Forbes. “The algorithm that MIT used solved—on average—Wordle’s puzzles with just 3.421 guesses, which they determined was better than audio or crane,” writes Kain. “And salet performed 1% better than Wordle Bot’s recent best word, slate, despite having the same letters.”
MIT researchers have found that “Salet” is statistically the best starting word for Wordle, reports Nicolas Vega for CNBC. “If you play Salet and you play intelligently you can assuredly win the game within five guesses,” says graduate student Alexander Paskov. “You don’t even need that last guess.”
Researchers from MIT are working with the Staten Island Performing Provider System to develop an algorithm that can predict who in the system is at risk for an opioid overdose, reports Maddie Bender for the Daily Beast. “In preliminary testing, Conte’s team and MIT Sloan researchers found their model was highly accurate at predicting overdoses and fatal overdoses, even with delays in the data of up to 180 days,” writes Bender.
MIT researchers have developed a new hardware that offers faster computation for artificial intelligence with less energy, reports Kyle Wiggers for TechCrunch. “The researchers’ processor uses ‘protonic programmable resistors’ arranged in an array to ‘learn’ skills” explains Wiggers.
Loh Down on Science host Sandra Tsing Loh spotlights Prof. Cathy Wu and graduate student Vindula Jayawardana and their work developing a new method for self-driving vehicles that would help minimize idling at red lights. “In their method, self-driving can be taught to minimize stops at red lights. To make this work, traffic lights and self-driving cars would have sensors. This would let them check in with each other on their surroundings,” says Loh.
Computer science lecturer Iddo Drori and his team have developed an artificial intelligence algorithm that can solve college-level math problems at a human level, reports Susan D’Agostino for Inside Higher Ed. “The model can also explain the solutions and generate new problems that students found indistinguishable from human-generated problems,” reports D’Agostino.
The MIT mini cheetah broke a speed record after learning to adapt to difficult terrain and upping its speed, reports Rienk De Beer for Popular Mechanics.