{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2603-24422","title":"OneSearch-V2: The Latent Reasoning Enhanced Self-distillation Generative Search Framework","arxiv_id":"2603.24422","date":"2026-03-25","proceeding":null,"authors":["Ben Chen","Siyuan Wang","Yufei Ma","Zihan Liang","Xuxin Zhang","Yue Lv","Ying Yang","Huangyu Dai","Lingtao Mao","Tong Zhao","Zhipeng Qian","Xinyu Sun","Zhixin Zhai","Yang Zhao","Bochao Liu","Jingshan Lv","Xiao Liang","Hui Kong","Jing Chen","Han Li","Chenyi Lei","Wenwu Ou","Kun Gai"],"abstract":"Generative Retrieval (GR) has emerged as a promising paradigm for modern search systems. Compared to multi-stage cascaded architecture, it offers advantages such as end-to-end joint optimization and high computational efficiency. OneSearch, as a representative industrial-scale deployed generative search framework, has brought significant commercial and operational benefits. However, its inadequate understanding of complex queries, inefficient exploitation of latent user intents, and overfitting to narrow historical preferences have limited its further performance improvement. To address these challenges, we propose OneSearch-V2, a latent reasoning enhanced self-distillation generative search framework. It contains three key innovations: (1) a thought-augmented complex query understanding module, which enables deep query understanding and overcomes the shallow semantic matching limitations of direct inference; (2) a reasoning-internalized self-distillation training pipeline, which uncovers users' potential yet precise e-commerce intentions beyond log-fitting through implicit in-context learning; (3) a behavior preference alignment optimization system, which mitigates reward hacking arising from the single conversion metric, and addresses personal preference via direct user feedback. Extensive offline evaluations demonstrate OneSearch-V2's strong query recognition and user profiling capabilities. Online A/B tests further validate its business effectiveness, yielding +3.98\\% item CTR, +2.07\\% buyer volume, and +2.11\\% order volume. Manual evaluation further confirms gains in search experience quality, with +1.37\\% in page good rate and +1.65\\% in query-item relevance. More importantly, OneSearch-V2 effectively mitigates common search system issues such as information bubbles and long-tail sparsity, without incurring additional inference costs or serving latency.","url_abs":"https://arxiv.org/abs/2603.24422","url_pdf":"https://arxiv.org/pdf/2603.24422","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2603.24422","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2603.24422"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/benchen4395/onesearch-family","reach":null}],"summary":{"unverified":6},"by_repo_kind":{"found_in_text":{"samples":6,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":6,"samples":[{"code_sha256_prefix":"f1628ae626696c59","entry":"composite_item_reward_func","repo":"benchen4395/onesearch-family","repo_kind":"found_in_text","path":"rlhf/train_grpo.py","file_url":"https://github.com/benchen4395/onesearch-family/blob/HEAD/rlhf/train_grpo.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"f1628ae626696c59"}},{"code_sha256_prefix":"7e7af3c25d8c53fc","entry":"concatenated_inputs","repo":"benchen4395/onesearch-family","repo_kind":"found_in_text","path":"rlhf/listwisedpo.py","file_url":"https://github.com/benchen4395/onesearch-family/blob/HEAD/rlhf/listwisedpo.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"7e7af3c25d8c53fc"}},{"code_sha256_prefix":"7e4675a4df245900","entry":"concatenated_inputs_sdpo","repo":"benchen4395/onesearch-family","repo_kind":"found_in_text","path":"rlhf/listwisedpo.py","file_url":"https://github.com/benchen4395/onesearch-family/blob/HEAD/rlhf/listwisedpo.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"7e4675a4df245900"}},{"code_sha256_prefix":"376fa049b3508a72","entry":"l2_normalize","repo":"benchen4395/onesearch-family","repo_kind":"found_in_text","path":"rq-opq/keyword_enhance.py","file_url":"https://github.com/benchen4395/onesearch-family/blob/HEAD/rq-opq/keyword_enhance.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"376fa049b3508a72"}},{"code_sha256_prefix":"de3af5a1e53bffd1","entry":"merger_query_item","repo":"benchen4395/onesearch-family","repo_kind":"found_in_text","path":"rq-opq/rq_dynamic.py","file_url":"https://github.com/benchen4395/onesearch-family/blob/HEAD/rq-opq/rq_dynamic.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"de3af5a1e53bffd1"}},{"code_sha256_prefix":"21ba0273539ebdcc","entry":"pad_to_length","repo":"benchen4395/onesearch-family","repo_kind":"found_in_text","path":"rlhf/listwisedpo.py","file_url":"https://github.com/benchen4395/onesearch-family/blob/HEAD/rlhf/listwisedpo.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"21ba0273539ebdcc"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CL","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}