arXiv cs.AI· Zhenqing Ling, Qirui Jiao, Daoyuan Chen, Cheng Chen, Xinji Mai, Haowen Gao, Zhuo Zhang, Xuchen Pan, Mengsu Ding, Jiaji Deng, Yuexiang Xie, Xiaoming Qin, Weirui Kuang, Zhaoyang Liu, Yaliang Li, Bolin Ding, Jingren Zhou·· 10 小时前AI 评分55
Piccolo:通过递归自我改进协同演化决策模型与数据
Piccolo: Co-Evolving Decision Models and Data through Recursive Self-Improvement
AI 导读
研究团队提出 Piccolo 系列多模态决策模型(0.8B 至 27B),通过 Agentic Data Engineering 与递归自我改进(RSI)让模型能力与训练数据协同演化。
正文 · 原文
Authors:Zhenqing Ling, Qirui Jiao, Daoyuan Chen, Cheng Chen, Xinji Mai, Haowen Gao, Zhuo Zhang, Xuchen Pan, Mengsu Ding, Jiaji Deng, Yuexiang Xie, Xiaoming Qin, Weirui Kuang, Zhaoyang Liu, Yaliang Li, Bolin Ding, Jingren Zhou
Abstract:We present Piccolo, a family of multimodal decision models (0.8B to 27B) that coevolves model capabilities and training data via Agentic Data Engineering and recursive self-improvement (RSI). The models predict probabilities over candidate answers for general-purpose text and visual judgments, agent harnesses, data workflows, and interactive games. Piccolo-27B tops the Decision Index Vision Full leaderboard with a score of 75.59, while Piccolo-9B and Piccolo-4B achieve 68.35 and 67.46, respectively. All three rank first on the Private benchmark in their respective model-size categories. Research agents analyze model behavior, construct training data, conduct experiments, and use the results to refine subsequent proposals. Semantic slices and a shared research memory connect these actions across the model family's development. With training recipes comprising hundreds of thousands of examples, data selection and feedback-driven refinement are central to the self-improvement process. Our large-scale RSI exploration yields three counter-intuitive insights for decision-model development: (1) Backbone-recipe co-adaptation outweighs raw model recency: earlier backbones can surpass newer generations on visual tasks under identical recipes; (2) Targeted data curation enables 4B-9B models to rival 27B baselines on specialized workflows; and (3) Continuous loss reduction can decouple from and even degrade discrete decision accuracy. Together with the released weights and tools from the Data-Juicer and AgentScope ecosystems, Piccolo provides an actionable blueprint for building frontier decision models and scaling data-driven RSI.
| Comments: | 22 pages, 13 figures, 11 tables. Project page and blog: this https URL |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2611.24651 [cs.CL] |
| (or arXiv:2611.24651v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.24651 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Daoyuan Chen [view email]
[v1]
Wed, 18 Nov 2026 15:53:39 UTC (1,777 KB)
来源:arXiv cs.AI · arxiv.org