[Submitted on 14 Jul 2025 (v1), last revised 18 Oct 2026 (this version, v4)]
Abstract:Scientific paper retrieval, particularly framed as document-to-document retrieval, aims to identify relevant papers in response to a long-form query paper, rather than a short query string. Previous approaches to this task have largely focused on abstracts, embedding them into dense vectors as surrogates for full documents and calculating similarity between them. Yet, abstracts offer only sparse and high-level summaries, and such methods primarily optimize one-to-one similarity, overlooking the dynamic relations that emerge across relevant papers during the retrieval process. To address this, we propose Chain of Retrieval (CoR), a novel iterative framework for full-paper retrieval. Specifically, CoR decomposes each query paper into multiple aspect-specific views, matches them against segmented candidate papers, and iteratively expands the search by promoting top-ranked results as new queries, thereby forming a tree-structured retrieval process. The resulting retrieval tree is then aggregated in a post-order manner: descendants are first combined at the query level, then recursively merged with their parent nodes, to capture hierarchical relations across iterations. To validate this, we present SciFullBench, a large-scale benchmark providing both complete and segmented contexts of full papers for queries and candidates, and results show that CoR significantly outperforms existing retrieval baselines. Our code and dataset is available at this https URL.
Submission history
From: Sangwoo Park [view email]
[v1] Mon, 14 Jul 2025 08:41:53 UTC (328 KB)
[v2] Sat, 1 Nov 2025 07:56:06 UTC (1,241 KB)
[v3] Sat, 21 Mar 2026 03:10:15 UTC (887 KB)
[v4] Sun, 18 Oct 2026 06:35:58 UTC (889 KB)
