Scientific American reporter Deni Ellis Béchard spotlights graduate student Alex Zhang for the publication’s special section featuring 28 rising young scientists. “The types of research that I want to work on are things that I think should be shared for the benefit of people in general,” Zhang says of his work aimed at improving AI user experience with recursive language models.
Writing for MassLive, Scott Kirsner highlights Gander Robotics, a startup co-founded by Sloan graduate student Michael Autery, that is “developing a torpedo-like device that is light enough to be tossed over the side of a ship — and designed to find a person in the water.” Autery, who served 15 years in the U.S. Navy, first pitched the idea for an “autonomous rescue swimmer” as part of an MIT entrepreneurship competition and notes: “if I had the same idea in a different place at the same time, I’m not sure it would’ve played out the same way. Cambridge is this very rich ecosystem for entrepreneurship.”
A new study co-authored by Sloan graduate student Anand Shah explores the growing use of AI across the legal system, reports Chris Stokel-Walker for Fast Company. “The pro se share of all civil cases has been 11% for quite some time,” says Shah. “And then in the post-AI world, we see it jumping all the way up to something like 18%.”
Nature reporter Rachel Fieldhouse spotlights graduate student Lauren 'Ren' Ramlan’s work integrating the video game Doom into her research. Ramlan used “Escherichia coli bacteria to display a few frames of Doom,” explains Fieldhouse. “She attached a fluorescent protein to the bacterial cells that could be turned on or off, making them act like black and white pixels on a screen. She then translated and compressed the first few frames of Doom into black-and-white versions that matched the plate growing the cells. Ramlan says the project shows what living things can be engineered to do, and that bacteria are probably not suitable for computing or displaying images.”
Graduate student Verena Bellscheidt has been awarded the Zonta Club of Wuaboag Valley Women in STEM award, reports The Republican. Bellscheidt is “dedicated to advancing the field [of theoretical physics] at the intersection of cosmology and particle physics while mentoring the next generation of female scientists.”
Postdoctoral Fellow Florian Galleri speaks with Associated Press reporter Sylvie Corbet about France’s move to align “its nuclear deterrent strategy more closely with European allies while keeping full control over any strike decision.” Galleri says “the strategic backing intended to integrate French nuclear deterrence into a collective European defense framework necessarily requires a degree of coordination and joint planning.”
Prof. Anna-Christina Eilers and postdoctoral associate Rohan Naidu speak with Scientific American reporter Rebecca Boyle about the discovery and study of Little Red Dots, mysterious, red spots that showed up in images from the James Webb Space Telescope. The dots, which astronomers dated to 600 million years after the big bang, “are in every single image the telescope takes,” says Naidu. “We have to find out about them if we want to tell a complete story about the early universe."
Graduate student Clarke Cyrus and his colleagues have developed “the Anemoia Device,” a physical machine that “uses a generative AI model to analyze an archival photograph, describe it in a short sentence, and, following the user’s own inputs, convert that description into a unique fragrance,” reports Grace Snelling for Fast Company. “Generative AI usually starts with a blank prompt,” says Clarke. “The dials replace that with a physical, easy-to-understand grammar. You’re not trying to ‘say the right thing’ to an algorithm, it’s more akin to tuning an instrument."
Researchers at MIT have “developed a new technique that enables large language models to learn new skills and knowledge without forgetting their past capabilities,” reports Ben Dickson for Venture Beat. “Their technique, called self-distillation fine-tuning (SDFT), allows models to learn directly from demonstrations and their own experiments by leveraging the inherent in-context learning abilities of modern LLMs,” explains Dickson. “Experiments show that SDFT consistently outperforms traditional supervised fine-tuning (SFT) while addressing the limitations of reinforcement learning algorithms.”
Using the James Webb Telescope, postdoctoral associate Rohan Naidu and his colleagues have captured a glimpse of Galaxy MoM-z14, which existed 280 million years after the Big Bang, and could provide clues as to what the universe was like during its infancy and how it has evolved over time, reports Passant Rabbie for Gizmodo. “We can take a page from archeology and look at these ancient stars in our own galaxy like fossils from the early universe,” says Naidu. “Except in astronomy we are lucky enough to have Webb seeing so far that we also have direct information about galaxies during that time.”
Postdoctoral Associate Fatima Husain speaks with National Geographic reporter Theo Nicitopoulos about the various theories used to explain how living creatures survived the Cryogenian era. “These surface settings could have enabled a diverse assemblage of life to persist and continue to evolve throughout the glaciations,” says Husain. “We keep learning more about how extreme the Cryogenian was… and that makes life persisting and diversifying dramatically after that, all the more amazing.”
Writing for The Boston Globe, graduate student Jonah Prousky highlights the benefits and drawbacks of social media bans, and ways to protect kids from online harms. “We likewise need a cultural reckoning on screen time for kids, since heavy use is associated with anxiety, depression, attention fragmentation, and sleep deprivation,” writes Prousky. “Australia demonstrated that a law can be used to get the ball rolling.”
Graduate student Stephen Casper speaks with Wired reporter Matt Burgess about the rise of “deepfake video abuse and its role in nonconsensual intimate imagery generation.” “This ecosystem is built on the back of open-source models,” says Casper. “Oftentimes it’s just an open-source model that has been used to develop an app that then a user uses.”
A new analysis conducted by postdoctoral associate Rohan Naidu and his colleagues has found evidence that suggests “little red dot” galaxies may contain baby black holes, reports Alex Wilkins for New Scientist. “In ordinary black holes, what you actually see with your eyes is the tip of the iceberg of the total energy that is coming out of the system, but the little red dots we now understand should really be thought of as these puffed-up black hole stars,” says Naidu. “It seems that most of their energy is coming out at these wavelengths that we see with our eyes, so what you see is what you get.”
Alumnus Jerry Lu and his colleagues have developed OOFSkate, an AI-powered app that can analyze a figure skater’s “jump height, rotation speed, airtime and even landing quality,” reports Dave Skretta for the Associated Press. “Our vision for the system is to automate the technical calling of the sport,” says Lu. “This manifests itself in a combination of using AI-assisted computer vision, but also the knowledge of figure skating, essentially taking out the stuff that should be judged without subjectivity.”