TTIC is hiring tenure-track faculty! Join a top #computerscience research environment with exceptional PhD students, strong research support, and a uniquely light teaching load. Apply by Dec 1: buff.ly/8HZoyKl

Friday, October 30th at 11:00am: Research at TTIC presents Svetlana Lazebnik of the University of Illinois at Urbana-Champaign @siebelschool with a talk titled "Exploiting 'Foundation Models' for Robotic Manipulation." Please join us in Room 530, 5th floor.

What a wholesome video from @TTIC_Connect. TTIC is an amazing and unique academic place.

Like Greg in this video + my brother (who was visiting me then) said, "This feels like a startup environment for academics." It is exactly how I felt during my two years as an RAP at TTIC!

What does it take to advance fundamental #computerscience? In our new video, hear from TTIC faculty and students and step inside our research spaces, from #ML and #robotics to natural language processing. Watch on YouTube: buff.ly/khGNEC1

"'From Utterances to Networks: Modelling Slang Adoption and Diffusion Across Subreddits' by Xiaoning Wang, Ted Underwood, and TTIC Researcher Zhewei Sun will be presented at #EMNLP2026. Tuesday, Oct. 27, 9:00–10:30AM, Hall H, Session 10.

"One Frozen Simulator Is Not Enough: Simulator Collapse in Multi-Agent RL" by Simon Yu and TTIC Professor @NickATomlin, et. al, will be presented at #EMNLP2026. Monday, Oct. 26, 9:00–10:30AM, Hall H, Session 5.

"Domain-Specific Jargon in Large Language Models" by TTIC Professor Zhewei Sun and Darin Keng will be presented as an oral talk at #EMNLP2026! Monday, Oct. 26, 11:00AM–12:30PM, F8, Session 6.

Congratulations to Nicholas Tomlin (@NickATomlin), Zhewei Sun, Chenxiao Yang (@chenxiao_yang_), and David McAllester for presenting their papers, posters, and talks at #EMNLP26. See the lineup of TTIC researchers and presentations: buff.ly/9GLAZX8

1/6 "We already have robots that can do really well on specialized tasks, but they struggle to do everyday tasks that humans take for granted in realistic, cluttered, real-world situations," said Osbert Bastani of @Penn at his TTIC Colloquium talk. Video: youtube.com/watch?v=v54F4b…

5/6 "And even if we could scale them, there's the question of how we'd do reinforcement learning to integrate these simulation environments into foundation model training," said Bastani of @Penn.

6/6 To address these challenges, Bastani described a 2-step solution: first, scale the design of sim environments using LLMs and then integrate these designs into embodied foundation model training. His team has mainly addressed step 1 and is working on step 2, said Bastani.

We’re thrilled to welcome 5 new faculty to TTIC this fall!

@BrianDuSell
Gary Hoppenworth
William Merrill (@lambdaviking)
@NickATomlin
@nikostsilivis

Learn more about their research: buff.ly/9U6i6ST

Monday, October 26th at 11:30am: TTIC Colloquium presents Elisabetta Cornacchia (@Elisabetta68885) of Bocconi University with a talk titled "Learning with Neural Networks: Temporal Correlations and Hidden Structure." Please join us in Room 530, 5th floor.

Friday, October 23rd at 11:00am: Research at TTIC presents Jingyan Wang with a talk titled "The Double-Edged Sword of Information: Revealed versus Hidden Lotteries in School Choice." Please join us in Room 530, 5th floor.

"Scaling Sim2Real Reinforcement Learning," a TTIC Colloquium talk by Osbert Bastani @obastani of @CIS_Penn, is now available to stream on YouTube: buff.ly/jX2NJOD

Congratulations to Professor Nikolaos Tsilivis @nikostsilivis, who received a Coefficient Giving grant to support his project, “Theoretical Foundations of Robust Autoregressive Learning as a Means of Designing Safe and Robust Models.” Read more: buff.ly/slwvXKO

1/6 LLMs with CoT reasoning are powerful, but can be incorrect, even unsafe. “You might like to train...a verifier to interface with these generators to make them more reliable, safe, aligned with your values, and so on," said Avrim Blum, President of TTIC youtube.com/watch?v=qyzuz6…

5/6 The question then is, said Blum: "If you have a bound on how many soundness mistakes (k) you are willing to make, what is the minimum number of total mistakes that you might have to make?"

6/6 "You can show that this gives a minimax characterization. There is a deterministic algorithm that guarantees to make at most this many mistakes given the bound on the # of soundness errors, and no algorithm can guarantee less," said Blum. Full talk: youtube.com/watch?v=qyzuz6…