HaMo Combines Experience Cards with Model Adaptation for Long-Horizon Web Agents
Zhen Bi · hf · 2026-10-28
Researchers from Huzhou Normal University, Alibaba, ZJU and Ningxia University propose HaMo, a harness-model collaborative framework for long-horizon continual learning in web agents.
- Distills trajectories into strategy-level experience cards for fast non-parametric retrieval and reuse
- Trains the model with an experience-conditioned continual learning objective to interpret, ground and execute retrieved strategies
- Uses a retrieval-coverage-based selection mechanism to pick compact, complementary trajectories for parameter updates
Experiments show HaMo outperforms harness-only or model-only baselines, achieving steady within-domain improvement and better cross-domain retention.
More from coding & agent
- Early users praise DeckWerk after weeks of using it for real talks — vincesitzmann · 2026-10-28
- Agentic video editing goes pro: senior editors 5x output, costs to collapse — dotey · 2026-10-28
- Software is going agent-first: sites, browsers, even a new Rust-like language built for agents — mobileraj · 2026-10-28
- HABIT-Bench: Best agent memory system scores just 33.61% at inferring user habits — SJTU · 2026-10-28
- Regehr's 'Piano Test': would you trust formally verified code with your life? — moyix · 2026-10-28
- Vibe-coded 'AI 浮生记' game satirizes AI startup life with compute black-market trading — vista8 · 2026-10-28