There are competing views on whether RL can genuinely improve base model's performance (e.g., pass@128). The answer is both yes and no, largely depending on the interplay between pre-training, mid-training, and RL. We trained a few hundreds of GPT-2 scale LMs on synthetic

I’m at NeurIPS and giving a keynote at the @imageomics workshop on Saturday! If you want to learn more about my new lab’s research be sure to drop by, and don’t forget to stick around for all the great talks and posters throughout the workshop! https://t.co/5df1CCphmM

Jianyang is a fantastic collaborator. He led BioCLIP 2 to a huge success, and is also a great mentor to younger students and larger teams (FinerCAM, SST). 10/10 recommend working with/hiring Jianyang!

Going to #NeurIPS2025 San Diego? Escape the conference for a couple hours with a morning bird walk in the trails of Balboa Park. 7 am on Thurs, Dec. 4.

Life update: I moved to silicon valley to tackle agents' biggest challenges: plasticity and reliability. Today's agents are smart but brittle. They lack plasticity (continual learning and adaptation) and reliability (stable, predictable behavior with bounded failures). These two

BioCLIP 2 -> #neurips25 Spotlight! AC's comment restores my faith in peer-review: "I recommend this work for spotlight due to its potential impact in a relatively underexplored area. The work provides a large scale curated dataset, a trained embedding space, and extensive

Computer Use: Modern Moravec's Paradox A new blog post arguing why computer-use agents may be the biggest opportunity and challenge for AGI. https://t.co/6fZfTdx710 Table of Contents > Moravec’s Paradox > Moravec's Paradox in 2025 > Computer use may be the biggest opportunity

🧪 Chemists spend many hours planning and replanning synthetic routes for a target molecule to avoid dangerous reactants and intermediates.☠️🚫 🤔 What if an AI agent could plan around them automatically—better and faster than human experts? 🔬 Constrained retrosynthesis

🎉 Excited to share that our paper EvoSchema: Towards Text-to-SQL Robustness Against Schema Evolution was accepted at VLDB 2025! 🚀 📢 Reminder: join us at VLDB 2025 in London! 🗓️ Sept 2 (Tue), 10:45 AM – 12:15 PM 📍 Room Wordsworth 4F 📄 https://t.co/ZNAav4ZtoX #VLDB2025 #LLMs

Welcome back, students! At @imageomics and the @ABCGlobalCenter, we’re kicking off the semester with big questions: How can AI help us better understand and protect life on Earth? #AIforNature #AIforGood @OSUengineering

🚀 Still have a chance to submit to @NeurIPSConf for our Multi-Turn Workshop! 🏆 Best Paper Awards 🎓 10-15 Registration Waivers for student authors 🎤 New panelist: @willccbb from @primeintellect! ⏳ Deadline is August 22—only 10 days left! 🎉 Thanks to our sponsor

🚀 Excited to share our #ACL2025 Findings paper: Explorer — a scalable pipeline that generates diverse web trajectories via exploration, powering generalist GUI agents with strong performance! 📄 https://t.co/KXSsdxXIFQ 🌐 https://t.co/PcYG8QsVqF #WebAgents #SyntheticData #LLM

As AI agents start taking real actions online, how do we prevent unintended harm? We teamed up with @OhioState and @UCBerkeley to create WebGuard: the first dataset for evaluating web agent risks and building real-world safety guardrails for online environments. 🧵

Remember “Son of Anton” from the Silicon Valley show(@SiliconHBO)? The experimental AI that “efficiently” orders 4,000 lbs of meat while looking for a cheap burger and “fixes” a bug by deleting all the code? It’s starting to look a lot like reality. Even 18 months ago, my own

I'm excited to bring the Imageomics workshop to NeurIPS 2025! Consider submitting your work on ai4ecology, ai4conservation and general ai4science--if you're using images to learn something about the natural world, chances are it's a good fit for the imageomics workshop!

🚨 Postdoc Hiring: I am looking for a postdoc to work on rigorously evaluating and advancing the capabilities and safety of computer-use agents (CUAs), co-advised with @ysu_nlp @osunlp. We welcome strong applicants with experience in CUAs, long-horizon reasoning/planning,

🔎Agentic search like Deep Research is fundamentally changing web search, but it also brings an evaluation crisis⚠️ Introducing Mind2Web 2: Evaluating Agentic Search with Agents-as-a-Judge - 130 tasks (each requiring avg. 100+ webpages) from 1,000+ hours of expert labor -

Are you at #CVPR2025? RoboSpatial Oral is today! 📅 June 14 (Sat) | 🕐 1:00 PM | 📍Oral Session 4B @ ExHall A2

📢 Imageomics showcases Biodiversity + AI at #CVPR2025! 🔬 Jenna Kline presents MMLA 🎤 Jianyang Gu on static segmentation + ViT explainability 📊 Ankit Upadhyay on animal re-ID 🐟 Fish-Vista dataset for aquatic species @ICICLE_AI #AI4Science Read more:

Heading to #CVPR2025 to present our Oral paper with @NVIDIARobotics! 📅 June 14 (Sat) | 🕐 1:00 PM | 📍Oral Session 4B @ ExHall A2 I’ll also be at the 3D-VLA/VLM and EVAL-FoMo 2 workshops presenting the same work. Come say hi!