{"slug":"learning-through","label":"Learning Through","item_count":3,"day_count":3,"source_count":2,"first_seen":"2026-10-19T08:07:35+00:00","last_updated":"2026-10-26T14:25:40+00:00","generated_at":"2026-10-28T00:13:55.337169+00:00","sources":["arxiv_agent_systems_research","arxiv_cs_ai"],"days":[{"date":"2026-10-19","items":[{"title":"ClueFlow: Learning Long-Horizon Search through Attributed Research-State Transitions","url":"http://arxiv.org/abs/2610.25980v1","source":"arxiv_cs_ai","type":"paper","summary_1line":"Long-horizon search requires agents to acquire relevant evidence and integrate it into an evolving research state. Existing methods maintain research states to guide exploration or assign turn-level progress signals,...","why_it_matters":"Matches feed focus: agent, eval.","sid":"736cb3afaf8cf4b3","published":"2026-10-19T08:07:35+00:00","editor_note":"Attributes search progress to research-state transitions instead of turn-level signals."}]},{"date":"2026-10-23","items":[{"title":"Lifelong Context Adaptation through Bidirectional Memory-Skill Learning","url":"http://arxiv.org/abs/2610.32295v1","source":"arxiv_cs_ai","type":"paper","summary_1line":"Lifelong LLM agents must learn from an expanding history of interactions while making each decision within a bounded context window. Memory makes past experience available, but the agent may still need to infer how to...","why_it_matters":"Matches feed focus: agent, eval.","sid":"110ac885ad3cb773","published":"2026-10-23T07:53:18+00:00","editor_note":"Couples memory with learned skills so decisions fit a bounded context."}]},{"date":"2026-10-26","items":[{"title":"When Fine-Grained Credit Assignment Helps and Hurts: Multimodal Agentic Reinforcement Learning through Video Editing","url":"http://arxiv.org/abs/2610.35497v1","source":"arxiv_agent_systems_research","type":"paper","summary_1line":"Reinforcement learning (RL) has shown promise for training multimodal agents to reason and use tools. However, it remains unclear when fine-grained credit assignment improves learning and when it fails to improve task...","why_it_matters":"Matches feed focus: agentic.","sid":"868490bacebdc880","published":"2026-10-26T14:25:40+00:00","editor_note":"Identifies conditions where fine-grained credit assignment fails to improve task performance."}]}],"editorial":{"tldr":"A run of arXiv papers on how agents learn from their own trajectories. ClueFlow credits search steps by research-state transitions, and a memory-skill paper tackles lifelong learning under a bounded context window.","stale":false,"whats_new":"A multimodal RL paper on video editing asks when fine-grained credit assignment helps agent training and when it fails to improve task performance.","why_it_matters":"Credit assignment and memory design determine how agents improve from experience, which affects how you build training signals and context management for long-horizon agents.","take_for_builders":"Before adding per-step rewards to an agent RL setup, run an ablation against outcome-only rewards on your own tasks.","beats":[{"kicker":"SEARCH","tone":"launch","headline":"ClueFlow credits long-horizon search by research-state transitions","sids":["736cb3afaf8cf4b3"]},{"kicker":"MEMORY","tone":"rising","headline":"Bidirectional memory-skill learning for lifelong agents","sids":["110ac885ad3cb773"]},{"kicker":"NOW","tone":"now","headline":"Fine-grained credit assignment helps and hurts in multimodal agentic RL","sids":["868490bacebdc880"]}],"open_questions":["Do ClueFlow-style state-transition rewards hold up beyond the search benchmarks in the paper?","Which task properties predict that fine-grained credit assignment will hurt rather than help?"],"generated_at":"2026-10-28T00:13:28+00:00"}}