Google DeepMind's Dream‑RSI: Recursive AI Cuts Search Cost 162×
Google and DeepMind present Dream‑RSI, a recursive self‑improvement AI that replays past attempts to find better strategies, slashing search calls by up to 162×.
cpaua1 min74Filter by tag
#reinforcement-learning
All tagscpaua1 min74
Recurrent Looped Transformer: Decoding with Growing Depth per Token
The Recurrent Looped Transformer adds a recurrent decoder expanding computation depth with sequence length, per‑token cost constant, aiding RL and SFT.
cpaua1 min112
Harness-1: A 20B Search Agent with Externalized Memory and State
Harness-1 is a 20B-parameter search agent that offloads search state to an external harness, enabling strong long-task performance and better transfer to new benchmarks.
cpaua2 min698RL_Envs_101: Create RL Training Environments Across Frameworks
RL_Envs_101 helps you build reinforcement learning environments in OpenEnv, OpenReward, Verifiers, NemoGym & more, with examples and model-aware setup.
cpaua1 min334