I'm at #ICRA2024 and will be presenting my paper titled "Plug in the Safety Chip: Enforcing Constraints for LLM-driven Robot Agents" (yzylmc.github.io/safety-chip/). Excited++ 🦾🤖
Ankit Shah
@ankitjs
Research Scientist at The Boston Dynamics AI Institute. Prev at Brown CS, Ph.D. from MIT. Making robots easy to program and deploy. ajshah.info
Cambridge, MA
Joined August 2013
- How can robots reuse learned policies to solve novel tasks without retraining? In our #ICRA2024 paper, we leverage the compositionality of task specification to transfer skills learned from a set of training tasks to solve novel tasks zero-shot
- Using formal systems to ensure diversity in evaluations. Making sure we are evaluating reasoning when claiming that we are evaluating reason. Hope to see more such analysis accompanying every claim.More than 50% of the reported reasoning abilities of LLMs might not be true reasoning. How do we evaluate models trained on the entire internet? I.e., what novel questions can we ask of something that has seen all written knowledge? Below: new eval, results, code, and paper.
- I am super excited to be a part of the @talkingrobotics series on Dec 7. I also highly recommend checking out the previous seminars especially if you're interested in a human-centered approach to robotics research.💫Understand how (temporal) reward abstractions help planning and task realization with @ankitjs from @theaiinstitute in our next episode. 🤖#ArtificialIntelligence 📅Date: (Thursday) 07 Dec 2023 - 09:30 PST | 12:30 EST | 18:30 CET ✍️Register here: rb.gy/n5wlop
- At @corl_conf this week. Excited to present our work on leveraging temporal reward abstractions for helping with robot learning at the LEAP workshop.leap-workshop.github.io
