← Back to all news
arXiv Computation and Language By AllSpark Team

AdaptEvo: Learning under Imperfect Supervision through Confidence-Adaptive Training with Knowledge Evolution

Read the original on arXiv Computation and Language →

AdaptEvo is a framework that learns under imperfect supervision by combining confidence‑adaptive policy optimization (CA‑GRPO) with evolving decision knowledge and evaluation rubrics. Its training module balances outcome and process rewards based on reference confidence, while its evolution module extracts reusable knowledge from recurring failures and refines process rubrics to catch overlooked errors. On an industrial multimodal content‑moderation dataset, AdaptEvo with Qwen3.6‑35B‑A3B achieves 61.9% exact‑label accuracy and 72.2% binary decision accuracy on in‑period tests, outperforming GRPO by 7.5 and 3.7 percentage points, and maintains gains on out‑of‑period tests even as GRPO performance declines.

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 Computation and Language.

Related stories

arXiv Computation and Language
Oct 9

AdaptEvo: Adaptive Agent Learning with Evolving Supervision

AdaptEvo is a framework for learning under imperfect supervision in rule-governed contextual decision tasks. It couples confidence‑adaptive policy optimization (CA‑GRPO) with an evolution module that synthesizes reusable decision knowledge from recurring failures and refines process rubrics. On an industrial multimodal content moderation dataset, AdaptEvo outperforms baseline GRPO and fixed reward mixtures, achieving higher exact‑label and binary decision accuracy both in‑period and out‑of‑period.

By Shijun Wan, Jiancong Xie, Hang Xu, Jin Duan, Qixiong Wang, Xi Xiang, Maofei Que, Yahui Liu, Zhongyu Wei, Mu Chuan
agentsmultimodal
More like this →
arXiv AI
Oct 15

TextualRL: Textual Reinforcement Learning for Context Management

TextualRL is a gradient‑free reinforcement learning framework that updates the textual context guiding large language model agents without altering their parameters. It treats groups of rollouts from the same task as units of critique, contrasting successful and failed runs to generate context patches, and validates these patches using held‑out rewards. Experiments on six benchmarks show that TextualRL boosts test accuracy by 24.4 and 15.3 percentage points for GPT‑5.5 and Qwen3.8‑27B, respectively.

By Tianyu Huai, Jie Zhou, Bo Mao, Qin Chen, Liang He, Xipeng Qiu
llmsagentsreinforcement-learningbenchmarks
More like this →