UTCS is a recognized leader in creating the scientific knowledge and practical technologies exemplifying the digital revolution that defines the 21st century.
@UTAustin is launching a new School of Computing in fall 2026! With Information and Statistics & Data Science, we’ll expand student opportunities, accelerate research, and strengthen pathways to high-impact careers and grad study.
Read more: utex.as/3OQFrIh
1/3 UT Austin Computer Science is hiring faculty in several areas, including Programming Languages! We look forward to receiving your applications and welcoming new colleagues!
cs.utexas.edu/faculty/recrui…
⟳ Computer Science, The University of Texas at Austi retweeted
Check out our series of work at ICML and IMPL showing that even frontier models do not truly possess generalizable understanding of programming language semantics
CosmicAI Researchers and team explored the question “Can large language models (LLMs) condition their reasoning on explicitly supplied formal rules even when those rules conflict with their training priors?” @jessyjlicosmicai.org/news/llmconditi…
⟳ Computer Science, The University of Texas at Austi retweeted
Personal update: My lab will move to @UTCompSci in Summer'27! I have multi Ph.D. openings in AI infra and efficient inference, along with summer internship opportunities. I'll also continue building @EagleCorp as co-founder. Grateful to my friends at UWaterloo for their support!
⟳ Computer Science, The University of Texas at Austi retweeted
Backed by a new three-year NSF grant, researchers Lorenzo Torresani, Kristen Grauman, and Amy Pavel are seeking to leverage advances in AI video understanding to make interactive coaching available to all.
Read more: bit.ly/3UlSWmy
⟳ Computer Science, The University of Texas at Austi retweeted
🗣️ @UTAustin Computer Science is hiring for 2026–27!
We’re seeking:
• Tenured/tenure-track CS faculty
• Joint Neuroscience/CS faculty
• Instructional faculty
Join the new School of Computing and help shape the future of computing. 🤘
Apply: apply.interfolio.com/194129
⟳ Computer Science, The University of Texas at Austi retweeted
🚨 We introduce the problem of persistent calibration for trustworthy, continually learning models and find substantial headroom even for confidence estimators that appear well-calibrated at a single point in time.
As a model gains more knowledge, we would like its confidence to track that change in knowledge, without requiring repeated re-calibration. To test how existing methods do, we evaluate persistent calibration on checkpoints of open models, asking whether confidence estimators trained on earlier checkpoints can generalize to later ones, whose knowledge has changed. To measure the dependence of confidence on knowledge, we evaluate calibration on knowledge contrast sets: subsets containing questions that one checkpoint answers correctly and another checkpoint answers incorrectly, reflecting a change in knowledge.
Key findings:
1️⃣ Both inference-time and fine-tuning methods fall short on contrast-set calibration, even when they are well-calibrated on the full set. We quantify the headroom with oracle methods trained on future checkpoints – we call this an oracle because persistent calibration requires learning to estimate confidence in a way that generalizes to future checkpoints, whereas the oracle trains directly on them.
2️⃣ Multi-checkpoint training helps in both the non-oracle and oracle settings, suggesting an avenue for identifying confidence features that remain robust across changing knowledge.
3️⃣ Our analysis shows that two confidence adapters trained on different checkpoints can perform similarly on one checkpoint, but extremely differently on another checkpoint. This suggests that there is a vast space of possible confidence functions that are well-calibrated on a given checkpoint, out of which only some rely on meta-knowledge features that would generalize to other checkpoints, presenting a challenge for persistent calibration.
⟳ Computer Science, The University of Texas at Austi retweeted
Extremely excited to share our work in blogpost version. We study the conditions under which LLM agents develop their own languages, which we can’t understand, and introduce GlossoGen, a framework for doing this kind of research.
More details in the post and in the paper 👇
Much of how we oversee AI agents rests on one fragile assumption: that we can understand what they say to each other.
Read our blog post about GlossoGen: schmidtsciences.org/glossoge…
⟳ Computer Science, The University of Texas at Austi retweeted
I'm incredibly excited to lead the new NSF Center for Human and Robot Co-Adaptation!
It is one of three new $30M, 5yr Centers launched under the @NSF STC program, and is led by @UTAustin , with partners @IUBloomington , @MIT , @TuftsUniversity , @UUtah , and @Yale
A new $30 million @NSF award will establish the Center for Human and Robot Co-Adaptation at UT Austin.
Led by Texas researchers, the new center will advance groundbreaking research into how humans and robots can learn to live and work together.
Another bold step forward for research at Texas. 🤘
Read more: utex.as/4zIvi3w
⟳ Computer Science, The University of Texas at Austi retweeted
When do LLM agents develop new languages that we can’t understand?
Lots of recent news about this, based mostly on anecdata from a single run. We study language emergence more rigorously, finding key factors like LLM strength, access to scratchpad messages, and pressure for efficiency.
Studying the languages themselves, we find they are morphologically productive, compositional, and can be transmitted to new agents, including agents backed by weaker models, even ones not able to develop language on their own.
To study language emergence systematically, we developed a new platform, GlossoGen, which lets us design controlled, sandboxed multi-agent scenarios with different initial conditions and dynamics. We instantiate one such scenario and use it to study open and closed-weight models across many runs.
Key takeaways:
1️⃣ Sufficiently strong models, under pressure to communicate efficiently and with access to a postmortem scratchpad, develop new languages.
2️⃣ Languages are compositional and morphologically productive.
3️⃣ Languages can be transmitted to new learners who observe them being used without seeing their construction.
4️⃣ Even models that are not strong enough to construct languages can learn to use them. Agents take an active role in learning languages, with new agents repairing failed conversations via targeted queries.
More details in our paper below, including implications for safety/monitorability, cumulative cultural evolution, and linguistics.
🧵👇
⟳ Computer Science, The University of Texas at Austi retweeted
Very excited to share that our CRL-VLA paper has won the outstanding paper award at the Reinforcement Learning Conference (also best paper at the ICRA RL4IL workshop)! The full list can be found at rl-conference.cc/RLC2026Awar….
This is my first time attending @RL_Conference, and RLC has already became my favorite conference (alongside CoRL). The community, the great review process, the hot takes... Everything is amazing!
Congratulations to all co-authors: @jayjshim@ChenTangMark@yoonchangsung@cranialxix@PeterStone_TX@RobobertoMM
VLA models are capable generalists. But can they continually self-improve?
Such Continual Reinforcement Learning (CRL) problems are traditionally considered very challenging.
Surprisingly, we found that with the right setup, the simplest CRL recipe can work really well!
arxiv.org/abs/2603.11653