🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Come say hi at poster spot #124!

Looking forward to speaking at CoLLAs this year and making my first pilgrimage back to McGill as a non-student! 😊

Deep learning may be hard, but deep un-learning is even harder. 💪 How do we efficiently remove the influence of specific training examples while maintaining good performance on the remainder? Announcing NeurIPS Unlearning Competition 📢 Submit your best ideas!🏆

Agents equipped with neural networks gradually lose the ability to learn from new experiences. In the #NeurIPS2023 spotlight done with Google DeepMind, we carefully analyze the phenomenon in RL and offer a way to make agents more compute-efficient. arxiv.org/abs/2305.15555 🧵

I wrote a blog post about generalization in deep learning last year and then ChatGPT happened and it felt a bit antiquated. But given that 90's jeans are back in style, maybe 70's learning theory results can be too? open.substack.com/pub/facilisdes…

Registration for OATML’s Virtual Open Day closes tomorrow. If you are interested in joining for a PhD, we encourage you to register and join us 22 Nov from 4-5:30GMT! Sign-up: docs.google.com/forms/d/e/1FAI… Group members attending:

Figured out that they silently updated the webpage to remove authors from the highlighted line when I went to post a response telling reviewers that we'd updated the paper pdf and the comment functionality was gone 🙃

CoLLAs is one of my favourite conferences! Super collegial, focused, and friendly vibe -- and it's also on the way to the beach from Vienna if you're extending your ICML trip 😎

At #NeurIPS2023 and interested in the science of deep RL and compute-efficient agents? Clare Lyle will present our spotlight on plasticity loss today (Thu Dec 14) at 11:45AM CST! Great Hall & Hall B1+B2 (level 1) #1500 nips.cc/virtual/2023/p… Paper: arxiv.org/abs/2305.15555

Half a page, half a page, half a page over. Into the valley of Appendix rode the nuanced discussion of related work and limitations

📢We are thrilled to announce our first #CoLLAs2024 keynote speaker: Emtiyaz Khan

📢Mixtures of Experts unlock parameter scaling for deep RL! Adding MoEs, and in particular Soft MoEs, to value-based deep RL agents results in more parameter-scalable models. Performance keeps increasing as we increase number of experts (green line below)! 1/9

Roses are Red, Violets are Blue; Don’t Forget to Submit, (Tomorrow) on OpenReview!

Violets are blue Roses are red Missed ICML? Try CoLLAs instead :)

I am really excited to reveal what Google DeepMind's Open Endedness Team has been up to 🚀. We introduce Genie 🧞, a foundation world model trained exclusively from Internet videos that can generate an endless variety of action-controllable 2D worlds given image prompts.

Presented "Disentangling the Causes of Plasticity Loss in Neural Networks" today CoLLAs 2025 work led by Clare Lyle with Zeyu Zheng, Hado van Hasselt, James Martens, Razvan Pascanu, Will Dabney, and myself! Slides: tinyurl.com/2mbfbt6d +Paper: tinyurl.com/mrxcx92h🐨

📢 I'm very excited to release AgarCL, a new evaluation platform for research in continual reinforcement learning‼️ Repo: github.com/machado-resear… Website: agarcl.github.io Preprint: arxiv.org/abs/2505.18347 Details below 👇