Marc Lanctot

sharky6000.bsky.social

11 May 2025 08:37

A bit of an old paper but I'm still excited about it, I've never posted about it, and we just released code for the main algorithm: Population RL (PopRL). 👇 Population-based Evaluation in Repeated Rock-Paper-Scissors as a Benchmark for Multiagent Reinforcement Learning, accepted (TMLR '23) 🧵 1/N

Glen Berseth

glenberseth.bsky.social

11 May 2025 08:37

Generalization across plans is a difficult, data-intensive process. Here, I cover learning representations to allow plans to generalize. I make connections between system identification, sim-to-real, meta reinforcement learning, and in-context learning and each method's limitations and generality.

Carlo Pinciroli

carlo.pinciroli.net

11 May 2025 08:37

Best week ever, with Josh Bloom defending his PhD dissertation titled "Global State Prediction for Multi-Agent Reinforcement Learning in Collective Transport." Congratulations! Many thanks to Dr. Nicola Bezzo, Dr. Kevin Leahy, and Dr. Guanrui Li for your time and detailed feedback!

Chris Paxton

cpaxton.bsky.social

11 May 2025 08:37

Reinforcement learning promises us AI systems that can generate their own data and self-improve. It's a powerful tool, but will it solve all our data problems? No. Here's my write up about RL and when it's useful: itcanthink.substack.com/p/the-limits...

Sung Kim

sungkim.bsky.social

24 April 2025 15:26

Awesome RL Reasoning Recipes ("Triple R") A curated collection covering models, datasets, reward designs, optimization methods, hyperparameters, empirical findings, theoretical insights, and everything about reasoning with reinforcement learning. github.com/TsinghuaC3I/...