基於工作流程歸納的多輪代理型科學文獻搜索

摘要

科學文獻搜尋往往不僅需要透過單一查詢來檢索論文：使用者的意圖往往不明確、依賴個人偏好，並在互動過程中不斷演變。現有的搜尋代理通常依賴固定流程或僅以隱性語言進行推理，導致其搜尋策略難以控制、檢視與調整。我們提出 PaperPilot，這是一個多輪文獻搜尋代理，將科學搜尋視為工作流程歸納。給定一篇錨定論文與使用者查詢後，PaperPilot 會建構一個可執行的論文搜尋運算子有向無環圖（DAG），包含關鍵字搜尋、引用擴展、篩選、評分、重新排序與證據提取。接著，使用者回饋被用來優化查詢內容與工作流程本身。我們透過監督式工作流程模仿與受控工作流程損壞下的偏好最佳化來訓練 PaperPilot。實驗結果顯示，在多輪互動情境下，PaperPilot-9B 相較於基礎的 Qwen3.5-9B 工具集代理，其 Hit@5 從 58.0 提升至 77.0，MRR 從 47.5 提升至 59.4，nDCG@10 從 26.8 提升至 32.5，同時工作流程執行錯誤率從 9.5% 降至 0%。這些結果表明，明確且可編輯的搜尋工作流程為將文獻搜尋代理與複雜的科學意圖對齊，提供了一個有效且可控的介面。

English

Scientific literature search often requires more than retrieving papers from a single query: users' intents are underspecified, preference-dependent, and evolve through interaction. Existing search agents typically rely on fixed pipelines or implicit language-only reasoning, making their search strategies difficult to control, inspect, and refine. We introduce PaperPilot, a multi-turn literature search agent that frames scientific search as workflow induction. Given an anchor paper and a user query, PaperPilot constructs an executable DAG of paper-search operators, including keyword search, citation expansion, filtering, scoring, reranking, and evidence extraction. User feedback is then used to refine both the query and the workflow itself. We train PaperPilot with supervised workflow imitation and preference optimization over controlled workflow corruptions. Experiments show that PaperPilot-9B improves over the base Qwen3.5-9B toolset agent under multi-turn interaction, increasing Hit@5 from 58.0 to 77.0, MRR from 47.5 to 59.4, and nDCG@10 from 26.8 to 32.5, while reducing workflow execution errors from 9.5% to 0%. These results show that explicit, editable search workflows provide an effective and controllable interface for aligning literature search agents with complex scientific intent.