SPADE Paper: Single Model Self-Implements Environment Design and Agent Self-Play
_AndrewZhao · x · 2026-08-21
Natasha Jaques et al. release the SPADE paper, proposing recursive self-improvement via multi-agent RL and Unsupervised Environment Design (UED). The framework uses a single LLM to act as both an "Environment Designer" (building multi-turn RL training environments using the Gym step()/reset() API) and a "Reasoning Agent" that learns to solve them. The Designer optimizes task difficulty by maximizing the Agent's regret, achieving automatic scaling of training environments.
Related event: SPADE: Single-Model Self-Play Writes Its Own Training Environments(13 posts)→
More from Research
- MIT PhD Sharut Gupta joins DeepMind; Isola hails her representational alignment defense talk — phillip_isola · 2026-10-16
- 36 primitives, just 10 parameters each: a remarkably precise T-pose body fit — letandrewcook · 2026-10-16
- Interp researcher pushes back: if you can't simulate a neuron, you haven't explained it — nabla_theta · 2026-10-16
- MLP-Splatting (ECCV 2026) code released: MLPs as light-field primitives instead of Gaussians — AjdDavison · 2026-10-16
- Four-page solution to subcubic APSP builds intuition for Anthropic-AI-aided breakthrough — gautamcgoel · 2026-10-16
- ATAS: Zhejiang-led team builds world-state memory framework for consistent agentic video generation — CSProfKGD · 2026-10-16