rohit It’s likely not about divining a smarter training set. Need fundamental advances in learning to get vastly more efficient scaling, e.g. going beyond next token prediction, or Kolmogorov-Arnold Networks, or something such.

Back on April 1st I posted my three laws of robotics. Here are my three laws of AI. 1. When an AI system performs a task, human observers immediately estimate its general competence in areas that seem related. Usually that estimate is wildly overinflated. 2. Most successful AI

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

Solving new tasks zero shot using prior experience in a related task! Find Jason Liu @RSS at #ICRA2024

Enforcing safety constraints with an LLM-modulo planner. Presented by Ziyi Yang at #ICRA2024

Ever wonder if LLMs use tools🛠️ the way we ask them? We explore LLMs using classical planners: are they writing *correct* PDDL (planning) problems? Say hi👋 to Planetarium🪐, a benchmark of 132k natural language & PDDL problems. 📜 Preprint: arxiv.org/abs/2407.03321 🧵1/n

RL in POMDPs is hard because you need memory. Remembering *everything* is expensive, and RNNs can only get you so far applied naively. New paper: 🎉 we introduce a theory-backed loss function that greatly improves RNN performance! 🧵 1/n

We will hear from an amazing line of speakers at our #RSS2024 workshop on robotic task specification tomorrow at 2 PM (CEST) in Aula Hall B Hadas Kress-Gazit 🪷, Peter Stone, Yonatan Bisk, David Abel, Cédric More details at: sites.google.com/view/rss-tasks…

All the talk recordings are available at youtube.com/playlist?list=…

Jason Liu @RSS will present their IJCAIconf survey paper on robotic language grounding. Please check out his talk (8/8 11:30) if you are at #IJCAI2024 In colab w/ VPC, Raymond Mooney from UT Austin, Stefanie Tellex from Brown CS, David Watkins from The AI Institute

My (pure) speculation about what OpenAI o1 might be doing [Caveat: I don't know anything more about the internal workings of o1 than the handful of lines about what they are actually doing in that blog post--and on the face of it, it is not more informative than "It uses Python

Evaluation in robot learning papers, or, please stop using only success rate a paper and a 🧵 arxiv.org/abs/2409.09491

Recent results like Apple’s show that LLMs (even o1) flub on reasoning with simple changes to problems that shouldn’t matter. A consensus is building that it shows they are “just pattern matching.” But that metaphor is misleading: good reasoning itself can also be framed as “just

How can robots understand spatiotemporal language in novel environments without retraining? 🗣️🤖 In our #IROS2024 paper, we present a modular system that uses LLMs and a VLM to ground spatiotemporal navigation commands in unseen environments described by multimodal semantic maps

I like how they use the infinite rotation to make planning easier. Taking advantage of how your humanoid doesn't need to be human

Friends, need your help. Antariksh Bothale, a senior from IIT B has launched an incredible project of organizing all Sanskrit literature in one place, in a user-friendly manner. The service is free, not-for-profit, created purely out of passion. Media coverage will go a long way in

Ok I’ve read it now and as I expected the complaints about it are ill founded. It seems fine. I think it convincingly shows a lot of what I’ve been saying about these thinking models.

Wow! The core finding in the much-maligned Apple paper from Parshin Shojaee et al – that reasoning models generalize poorly in the face of complexity – has been conceptually replicated three times in three weeks. C. Opus sure didn’t see that coming. And a lot of people owe Ms.

Yale Philosophy offers a course on “Formal Philosophical Methods” — a broad introduction to probability, logic, formal semantics, etc. Instructor Calum McNamara has now made all materials for the course (78 pages) freely available static1.squarespace.com/static/6255ffe…

A new handheld interface from MIT gives anyone the ability to train a robot for tasks in fields like manufacturing. The versatile tool can teach a robot new skills using one of three approaches: natural teaching, kinesthetic training, & teleoperation: bit.ly/4nTAw6F