uper happy to be one of the organizers of this, eurips.cc, now an option to officially present your papers in Europe and save greenhouse gas!!!!! Please repost and spread the word!!! Please repost!!!!

very proud that my work on multi-agent debate for misinformation detection won best paper award at the ICML Conference CFAgentic workshop! check it out on arxiv: arxiv.org/abs/2410.20140 v grateful to all my co-authors and the support from BBC Research & Development 🥳

#ICML2025 PCs weigh-in on hidden LLM prompts

this seems a bit like punishing the victims rather than perpetrators

🚨 [Call for Papers] SEA Workshop @ NeurIPS 2025 🚨 📅 December 6, 2025 | 📍 San Diego, USA 🌐: sea-workshop.github.io Environments are the "data" for training agents, which is largely missing in the open source ecosystem. We are hosting Scaling Environments for Agents (SEA)

Finally, a good modern book on causality for ML: causalai-book.net by Elias Bareinboim. This looks like a worthy successor to the ground breaking book by Judea Pearl which I read in grad school. (h/t Joshua Safyan for the ref).

We are excited to announce the MARS Multi-Agent Embodied Intelligence Challenge 🎉, which will be held at the NeurIPS 2025 SpaVLE Workshop! Co-organized by Shanghai Jiao Tong University, University of Oxford, The University of Hong Kong, UC San Diego, and other international

The speed of Everlyn video generation for this level of photorealism/hyperrealism is astonishing. Congratulations to Ser Nam CY Harry Yang and team on this combined research + engineering marvel 👏

🚨 Paper Alert 🚨 ➡️Paper Title: Articulate3D: Zero-Shot Text-Driven 3D Object Posing 🌟Few pointers from the paper 🎯Authors of this paper proposed a training-free method, “Articulate3D”, to pose a 3D asset through language control. 🎯Despite advances in vision and language

Sir, we built this. A RL environment for learning reasoning at scale. GitHub: github.com/camel-ai/loong HF dataset: huggingface.co/datasets/camel… We extracted seed datasets from sources like textbooks, code libraries like sympy, networkX, Gurobi (math programming lib), rdkit

not so often I do computer vision papers these days, great work from Ishi Deb

Excited to share our new work: “Learning to See Before Seeing”! 🧠➡️👀 We investigate an interesting phenomeno: how do LLMs, trained only on text, learn about the visual world? Project page: junlinhan.github.io/projects/lsbs/

A big thank you to the fantastic coauthors! philip, ioannis patras, Adel Bibi, Fazl Barez!

Please find many more results on 4 LLMs (across base models, instruction-tuned models, and reasoning models), and ablations in the paper! 📰 Project: james-oldfield.github.io/tpc 💻 Code: github.com/james-oldfield… 📄 Paper: arxiv.org/abs/2509.26238

Big congrats to my PhD student Junlin (Hans) Han, the Meta MSL team and philip ! “Learning to See Before Seeing” shines a light on visual priors from language pretraining and offers a practical recipe for vision-aware LLMs.

Adversarial Manipulation of Tool Selection As LLMs increasingly power agents that interact with external tools, tool use has become an essential mechanism for extending their capabilities. These agents typically select tools from growing databases or marketplaces to solve user

This project would not have been possible without amazing collaborators! Very grateful to my co-leads on the project Alesia Ivanova (who is an MSc student at Oxford!) and Charlie London, along with Jack Cai (Jack Cai), my advisors at Microsoft - Shital Shah (Shital Shah) and

Thanks to everyone who worked on this, particularly the leads Sumeet Motwani and Alesia, along with Jack Cai , Shital Shah, Riashat Islam, Christian Schroeder de Witt, and philip!

This is a joint work by amazing folks from Oxford including Alesia Ivanova, Sumeet Motwani, Charlie London, Jack Cai, Christian Schroeder de Witt, philip and Riashat Islam from Microsoft Research. Arxiv: arxiv.org/abs/2510.07312 Code: github.com/AlesyaIvanova/…

Work done with my awesome collaborators Iván Arcuschin Arthur Conmy Neel Nanda philip as part of the MATS Research