← Back to all news
arXiv AI By Yeonsung Jung, Trilok Padhi, Sina Shaham, Dipika Khullar, Joonhyun Jeong, Ninareh Mehrabi, Eunho Yang

Co-Evolving Agents: Learning from Failures as Hard Negatives

Read the original on arXiv AI →

The paper introduces a co‑evolving framework where a target agent improves by learning from its own failures, and a separate failure agent is trained to generate hard negative trajectories. These hard negatives, derived from plausible but incorrect attempts, help the target agent better distinguish successful behavior from subtle errors. Experiments on online shopping, scientific reasoning, and interactive SQL querying show a 5.7% average reward improvement over baseline methods.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

Related stories

arXiv AI
2d ago

Survival is the Only Reward: Sustainable Self-Training Through Environment-Mediated Selection

arXiv:2601.12310v2 Announce Type: replace Abstract: Self-training systems often degenerate due to the lack of an external criterion for judging data quality, leading to reward hacking and semantic dr...

By Jennifer Dodgson, Alfath Daryl Alhajir, Michael Joedhitya, Akira Rafhael Janson Pattirane, Surender Suresh Kumar, Joseph Lim, C. H. Peh, Adith Ramdas, Steven Zhang Zhexu
agentsefficiencydata-engineering
More like this →
arXiv AI
Sep 2

HarnessEvolve: Learning from Reference Trajectories for Reliable Agent Self-Evolution

HarnessEvolve is a self‑evolving framework that improves agent harnesses—prompts, skills, tools, and execution logic—by learning from reference trajectories. It separates execution, evaluation, optimization, and gating into independent modules, addressing credit assignment failure, shortcut learning, and catastrophic forgetting. The approach uses reference trajectories to extract error signals, applies quality and performance gates to candidate updates, and validates updates on held‑out data, consistently outperforming state‑of‑the‑art baselines across diverse benchmarks.

By Wen Jiang, Mingmin Chu, Yimeng Tian, Qianxin Zhang, Haofei Yang, Rui Yang, Yang Liu, Tao Lv, Fangming Li
agentsbenchmarks
More like this →