Proposed to cut number of people involved in NSF activities by 70%. We are literally on the chopping board. Call your reps.

The bottom line with the proposed cuts to NSF is to cut the number of working scientists and disappear opportunities to train new scientists. This is a PROPOSED budget to Congress. Call your reps.

We now have a whole YouTube video explaining our MINDcraft paper, check it out! youtu.be/MeEcxh9St24

“Stop treating ‘AGI’ as the north-star goal of AI research,” has been accepted to #ICML2025! arxiv.org/abs/2502.03689

The future of embodied AI revolves around *collaborative* multi agent scenarios that need natural language communication, task delegation, resource sharing, and more ⛏️ Here are MINDcraft and MineCollab, a simulator and benchmark purpose built to enable research in this area!

It's *PO'clock, this time IRPO In-Context Ranking Policy Optimization! An RL algorithm inspired by trad retrieval that trains agents to more effectively use lists of documents in context for better multi-hop {QA, agentic tasks, and more}!

Human designed text games like Zork still pose a huge challenge to modern SoTA LLMs, but why? This analysis breaks the problem down into various reasoning problems - which even claude-3.7-thinking can’t solve! Check the paper for more 😎

👀 soon youtu.be/iDJ6GrHNoDs

They made me do video 😬 but for a good reason! We are launching an iOS app–it runs OLMoE locally 📱 We're gonna see more on-device AI in 2025, and wanted to offer a simple way to prototype with it App: apps.apple.com/us/app/ai2-o... Code: github.com/allenai/OLMo... Blog: allenai.org/blog/olmoe-app

I use Twitter more often than bsky bc the tempo and energy around research conversations remains better there, even as the site degrades in many ways and many accounts go dark. ML feels particularly “left behind” there, I started the MLSky feed in the hope that migration is a marathon not a sprint

Oh a full paper too: The Lessons of Developing Process Reward Models in Mathematical Reasoning arxiv.org/abs/2501.07301

Day 13: Triangles and nothing else. This is an evolutionary algorithm that recreates an image (mr darwin) by randomly drawing triangles in a population of images and selecting the one closest to the target, then repeat. #genuary #genuary13

Excited to start my #ARR #NLP reviews! I'll try my best and see if I can get 100% of my reviews to be 'great' this round. If you didn't see it already, ARR publishes how many of your reviews are considered to be 'great': stats.aclrollingreview.org Join me for the challenge :)

The paper I spoke about at the #NeurIPS2024 ML for Systems workshop is now on arxiv arxiv.org/abs/2412.15573! We use multi-agent RL to solve a classic combinatorial optimization problem (the Sequential Assignment Problem), by combining MARL with a classical polynomial time assignment algorithm. 1/3

Headed to #NeurIPS2024 this week! 🇨🇦🍁 I’m interested in embodied agents, multi-agent/human-AI collaboration, LLMs, reinforcement learning and these cute kitties :) Send me a DM if you want to connect and chat!

I'm not at #NeurIPS2024 this week but my students and collaborators @izzycw.bsky.social, Zack Ankner, and Seungone Kim will all be there presenting various papers. Go chat with them about everything from multi agent RL with LLMs, reward modeling, personalized RLHF, and more!

Really excited for this event! Come say hi!! 🎉🏳️‍🌈🏳️‍⚧️

🇨🇦 Attending my last NeurIPS as a PhD student! Presenting our recent work, Patient-Ψ, at a few workshops --- come by! Schedule in 🧵 I'm on the job market, looking for TT faculty roles & post-docs. DM if you'd like to chat (or invite me to a party 🥳)!

There is now a "Starter Pack" of UCSD CSE Folks: go.bsky.app/AT3p8Vq

I had a great time at SoCal NLP today!! Thank you everyone for the wonderful conversations and insightful talks 😺

Papers from our group! 🤓 - Queer culture in television: 2024.computational-humanities-research.org/papers/paper... - Acting in American film: naitian.org/once-more-wi... - Classification w/ LLMs in cultural analytics: 2024.computational-humanities-research.org/papers/paper...

There! I went for it! (Let me know everyone if you want me to add or remove you) go.bsky.app/CUuio7g

Good morning Bluesky! I will presenting this today at #EMNLP2024 at 2 pm in the Jasmine room 🎉 Come by and say hi if you are interested in learning more about cross cultural communication, LLM agents, and AI for games! 🎮

I guess I should cross post here too. I'm recruiting one (1) PhD student to work on multimodal embodied agents. VLMs + RL. Please apply to the UCSD CSE app by Dec 15 and mark my name as a faculty of interest. More lab info at pearls.ucsd.edu

Excited to share that I will be at #EMNLP2024 presenting our work Communicate to Play! 🕑Session 12 14:00 - 15:30 📍Jasmine - Lower Terrace Level Come chat about LLM agents, cultural considerations in NLP, and AI for games!