Hy-Memory-30B-A3B: Attributed Reinforcement Learning for Long-Term Memory Writing

Key points

• Existing frameworks typically place a general-purpose large language model behind framework-specific prompts.1
• We present Hy-Memory-30B-A3B, a compact long-term-memory writer trained across Mem0, MemOS, and Hy-Memory.1
• Memory On-Policy Distillation teaches framework-specific write protocols on student-generated states; Memory RL then optimizes downstream answerability using dense per-chunk probes, attributed cross-chunk rewards, over-storage controls, and chunk-level GRPO.1
• Carrying policy-written banks across windows keeps training aligned with deployment.1

Sources (1)

• [1]Hy-Memory-30B-A3B: Attributed Reinforcement Learning for Long-Term Memory Writing
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 04:46 PM

  External-memory systems extend language-model applications beyond a context window, but their quality depends critically on the model executing the write path: deciding what to retain, how precisely to represent it, and how new evidence should update persistent state.

  Existing frameworks typically place a general-purpose large language model behind framework-specific prompts.

Extractive summary: sentences quoted from the sources.