Darío Gil SM ’00 PhD ’03, an MIT EECS alum, member of the MIT School of Engineering Dean’s Advisory Council, and IBM Chair of the MIT-IBM Watson AI Lab, has been elected as chair of the National Science Board. He is the NSB’s first industry chair in over 30 years. nsf.gov/nsb/news/news_…

Today at 2:30 pm EDT: Tune in for the joint MIT School of Engineering and Schwarzman College of Computing Advanced Degree Ceremony as we celebrate the achievements of our graduates! Live webcast: bit.ly/SoE-SCC-2024

Looking for a specific action in a video? Massachusetts Institute of Technology (MIT) and MIT-IBM Watson AI Lab researchers developed a method that trains a machine-learning model to perform this task, known as spatio-temporal grounding, using only videos and their automatically generated transcripts. news.mit.edu/2024/ai-based-…

Inside MIT’s Schwarzman College of Computing: bit.ly/4aOEJ3F

Congrats on the best paper, AWQ at MLSys'24!

Congratulations Nancy Kanwisher @[email protected]!

A new MIT-IBM Watson AI Lab ML method converts visual representations into text captions of a robot's observations. Along with instructions, the technique can direct a robot through a multistep navigation task without needing to train with costly visual data. news.mit.edu/2024/researche…

Professor and physician Collin Stultz wants to help heart patients everywhere by applying machine-learning techniques to cardiovascular medicine. The goal: “more equitable and fairer care for everyone,” he says. mitsha.re/qTnK50SkVwc

The School of Engineering has awarded Professor Leslie Kaelbling of @miteecs the 2024 Ruth and Joel Spira Award for Excellence in Teaching. The award is given annually to faculty members that embody the tradition of high-quality engineering education at MIT.

A picture is worth a thousand words, but can a LLM get the picture if it has never seen images before? 🧵 MIT CSAIL researchers quantify how much visual knowledge LLMs (trained purely on text) have. The visual aptitude of the language model is tested by its ability to write,

It's fantastic to hear that DiffTactile won the Best Paper Award at the RSS Tactile Sensing for General Purpose Robot Learning workshop! Difftactile is also integral to our Genesis Simulator. github.com/Genesis-Embodi… Great work from Zilin Si Gu Zhang and Elgce

MIT researchers and their colleagues examined how LLMs fare with variations of different tasks, putting their memorization and reasoning skills to the test, and seeing how they performed in familiar scenarios vs novel ones. news.mit.edu/2024/reasoning…

To estimate the reliability of foundation models, MIT and MIT-IBM Watson AI Lab researchers developed a technique to assess the consistency of representations similar models learn about the same test data point, before choosing a model for a downstream task. news.mit.edu/2024/how-asses…

A new machine-learning model can identify the stage of ductal carcinoma in situ (DCIS), potentially helping clinicians assess the current breast tumor stage, likeliness to progress to invasive cancer, and ultimately help in reducing overtreatment. news.mit.edu/2024/ai-model-…

Thermometer, a method for calibrating a large language model from MIT and MIT-IBM Watson AI Lab researchers, could help pinpoint situations where a LLM is overconfident and enable the model to produce better-calibrated responses on tasks it has not seen before. news.mit.edu/2024/thermomet…

MIT researchers developed MAIA, an automated, multimodal agent that can iteratively design experiments to better interpret various components of other models, for example, to understand particular neuron behaviors. news.mit.edu/2024/mit-resea…

Collin Stultz from MIT EECS and MIT IMES, is investigating how patients with cardiovascular disease could benefit from machine-learning techniques. betterworld.mit.edu/spectrum/issue…

Building on a landmark algorithm, researchers have proposed a way for a quantum computer to break certain cryptographic systems. Their quantum factoring circuit requires less memory and is more tolerant to noise than prior versions. mitsha.re/UuW350T4TKB

Super excited to share that I've defended my PhD MIT CSAIL! 🎓 I want to extend my immense gratitude to my advisor, Jacob Andreas, and my thesis committee — Constantinos Daskalakis, Gabriele Farina, and Roger Levy. Huge thanks to Noam Brown, Adam Lerer, Emily Dinan, Mike Lewis, and

A new algorithm solves complicated partial differential equations by breaking them down into simpler problems, potentially guiding computer graphics and geometry processing. mitsha.re/JbKw50TaN6b