Interactive agents can turn experience into reusable skills, yet existing self-evolving skill libraries primarily improve by accumulating new knowledge. Failures may lead to new skills, while previously stored skills are less often revisited as new evidence arrives. However, growth alone does not ensure reliability, as...
Jie Ma, Zhi-Peng Qian, Yu-Fei Ma et al.· 0 citations
Multi-agent systems enable complex reasoning and tool use by coordinating agents that divide roles and refine candidate solutions. Existing methods typically update individual agent responses or treat a complete trajectory as one training example. However, these methods may produce misleading policy updates because the...
Sheng-Tian Yang, Zi-Yun Xiong, Yu Li et al.· 0 citations
This work proposes a Causal auto-Bidding method based on a Diffusion completer-aligner framework, termed CBD, which achieves superior performance on large-scale auto-bidding benchmarks, but also delivers significant improvements on an online advertising platform, including a 2.0% increase in target cost.
Ye-Wen Li, Jingtong Gao, Peng Jiang et al.· Proceedings of the 32nd ACM...· 0 citations
This paper introduces \methodname, the first self-evolving search framework that eliminates difficulty-based proposer rewards and directly optimizes for evidence necessity relative to shortcut contexts, and shows that Dr. Free outperforms prior data-free search agents and the supervised baseline.
Zhi-Peng Qian, Zi-Han Liang, Yu-Fei Ma et al.· 0 citations
Auto-bidding is central to computational advertising, where strategies must maximize advertisers'conversion value under economic constraints. It has evolved from rule-based controllers to reinforcement learning and generative methods such as Decision Transformer (DT). Yet these methods increasingly mismatch the prevail...
Ye-Wen Li, Peng Jiang, Yi-Tian Li et al.· 0 citations
SRPO (Setwise Relative Policy Optimization), which treats the active set the minimal set of outputs consumed by one transition, as one multi-agent action, and unifies division of labor and joint co-evolution as actions with different set sizes.
Sheng-Tian Yang, Zi-Yun Xiong, Yu Li et al.· 4 citations
AgentBrew is proposed, an offline training framework that learns effective tool-use policies from a single batch of raw interaction trajectories, without task verifiers or iterative on-policy rollouts, and demonstrates that fine-grained offline learning can recover useful supervision from raw trajectories that filterin...
Zhiyi Lyu, Ye-Wen Li, Long-Tao Zheng et al.· 2 citations
This work proposes PRQ-KMeans, which removes the global-mean component, refines centroids with Top-k similarity-weighted updates, and replaces full-codeword subtraction with a projection residual that removes each representation's selected-centroid component.
Yunxiao Luo, Siyuan Wang, Ben Chen et al.· 0 citations
Hierarchical Residual Policy Optimization (HRPO), a post-training framework that converts item-level outcomes into dense, token-aligned learning signals for conservative token-wise improvement, is proposed.
Kaifeng Guo, Yiming Yang, Jingtong Gao et al.· Proceedings of the 32nd ACM...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
