{"ID":2849270,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.24698","arxiv_id":"2510.24698","title":"ParallelMuse: Agentic Parallel Thinking for Deep Information Seeking","abstract":"Parallel thinking expands exploration breadth, complementing the deep exploration of information-seeking (IS) agents to further enhance problem-solving capability. However, conventional parallel thinking faces two key challenges in this setting: inefficiency from repeatedly rolling out from scratch, and difficulty in integrating long-horizon reasoning trajectories during answer generation, as limited context capacity prevents full consideration of the reasoning process. To address these issues, we propose ParallelMuse, a two-stage paradigm designed for deep IS agents. The first stage, Functionality-Specified Partial Rollout, partitions generated sequences into functional regions and performs uncertainty-guided path reuse and branching to enhance exploration efficiency. The second stage, Compressed Reasoning Aggregation, exploits reasoning redundancy to losslessly compress information relevant to answer derivation and synthesize a coherent final answer. Experiments across multiple open-source agents and benchmarks demonstrate up to 62% performance improvement with a 10--30% reduction in exploratory token consumption.","short_abstract":"Parallel thinking expands exploration breadth, complementing the deep exploration of information-seeking (IS) agents to further enhance problem-solving capability. However, conventional parallel thinking faces two key challenges in this setting: inefficiency from repeatedly rolling out from scratch, and difficulty in i...","url_abs":"https://arxiv.org/abs/2510.24698","url_pdf":"https://arxiv.org/pdf/2510.24698v1","authors":"[\"Baixuan Li\",\"Dingchu Zhang\",\"Jialong Wu\",\"Wenbiao Yin\",\"Zhengwei Tao\",\"Yida Zhao\",\"Liwen Zhang\",\"Haiyang Shen\",\"Runnan Fang\",\"Pengjun Xie\",\"Jingren Zhou\",\"Yong Jiang\"]","published":"2025-10-28T17:51:50Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\"]","methods":"[\"LoRA\"]","has_code":false}