发表机构
Institute of Computing Technology, Chinese Academy of Sciences; University of Chinese Academy of Sciences; Peking University; DeepWisdom; Shanghai Artificial Intelligence Laboratory; Renmin University of China（中国科学院计算技术研究所; 中国科学院大学; 北京大学; 深度智慧; 上海人工智能实验室; 中国人民大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对现有LLM构思方法受静态工作流限制的问题，提出MindFlow框架，以概率思维超网建模图结构思维流，通过锦标赛相对排序优化，可生成可控优质研究思路，在多主题上表现优越。
AI 中文摘要
研究思路创新是科学进步的核心驱动力，但目前难以以可扩展且可控的方式生成与评估，这一挑战源于其固有的开放性与多目标属性，思路需兼顾新颖性、合理性与可行性。尽管近期基于大语言模型（LLM）的方法通过精心设计的提示词或智能体管线取得了一定进展，但受限于预定义的静态构思工作流。为解决这一局限，我们提出MindFlow框架，其将构思明确表述为思维超网（Mind Supernet）中由模块化思维算子构成的图结构思维流，采用概率思维超网建模。给定研究主题后，控制器会动态采样思维流以生成候选思路，该开放性问题通过基于锦标赛的相对排序优化，使控制器逐步倾向于更高质量的思维流。我们进一步引入评估协议，联合评估问题发现与问题解决，突破仅基于标题或摘要的判断。在不同主题上，MindFlow展现出作为明确、可控且可优化的研究思路创新工具的优越性。
英文摘要
Research idea innovation is a fundamental engine of scientific progress, yet it remains difficult to generate and evaluate in a scalable and controllable way. This challenge lies in its inherently open-ended and multi-objective nature, where ideas should balance novelty, plausibility and feasibility. While recent LLM-based approaches have made progress through carefully designed prompts or agent pipelines, they are constrained by predefined, static ideation workflows. To address this limitation, we propose MindFlow, a framework that explicitly formulates ideation as a graph-structured Flow in Mind, which is composed of modular thinking operators and modeled by a probabilistic mind supernet. Given a research topic, a controller dynamically samples thinking flows to generate candidate ideas. This open-ended problem is optimized using a tournament-based relative ranking, enabling the controller to progressively favor higher-quality thinking flows. We further introduce an evaluation protocol that jointly assesses problem finding and problem solving, going beyond title- or abstractonly judgments. Across diverse topics, MindFlow shows its superiority as an explicit, controllable and optimizable research idea innovator.
