{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DC2Q7SJTYY5GXHS6FUHLCUCUP2","short_pith_number":"pith:DC2Q7SJT","schema_version":"1.0","canonical_sha256":"18b50fc933c63a6b9e5e2d0eb150547ea377dff894cbae4711dd006985ab8963","source":{"kind":"arxiv","id":"2607.25344","version":1},"attestation_state":"computed","paper":{"title":"Reward Guided Decoding for Generative Recommendation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Jiawei Sheng, Lantao Hu, Liangliang Chen, Pengbo Xu, Ruochen Yang, Shijun Wang, Shuang Wen, Shuang Yang, Tingwen Liu, Wenwu Ou, Xiaoyu Zhang, Youfeng Zheng, Yusheng Huang, Zhaojie Liu","submitted_at":"2026-07-28T06:46:30Z","abstract_excerpt":"Generative recommendation formulates recommendation task into an SID sequence autoregressive generation paradigm, but the decoding process is often dominated by generation likelihood. This may conflict with real-world business objectives, where high-value candidates can receive low generation probability and be pruned early during beam search. Existing reranking or training-time alignment methods either intervene too late or require costly model retraining when business preferences change. To this end, we propose \\textbf{R}eward \\textbf{G}uided \\textbf{D}ecoding, named \\textbf{RGD}, a controll"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2607.25344","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-07-28T06:46:30Z","cross_cats_sorted":[],"title_canon_sha256":"f06cefb8612d19a784aa1ed9b1fd41d93a9633050af8d7948be9d4285dfe734a","abstract_canon_sha256":"3d3a1b81ddc2b5ce6bc6fc31aa4cfb91d8c80186c440423ec417d3c52ca106ac"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-29T01:25:10.402395Z","signature_b64":"51ojt4FP8CJ/+0KpE0d8Qa6zPVfeWF1Qi2/jCFmqmRmmN165dR2xxX4uuccj/1NEdkCIWTYv5FxeIcAIsxQZCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"18b50fc933c63a6b9e5e2d0eb150547ea377dff894cbae4711dd006985ab8963","last_reissued_at":"2026-07-29T01:25:10.401548Z","signature_status":"signed_v1","first_computed_at":"2026-07-29T01:25:10.401548Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reward Guided Decoding for Generative Recommendation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Jiawei Sheng, Lantao Hu, Liangliang Chen, Pengbo Xu, Ruochen Yang, Shijun Wang, Shuang Wen, Shuang Yang, Tingwen Liu, Wenwu Ou, Xiaoyu Zhang, Youfeng Zheng, Yusheng Huang, Zhaojie Liu","submitted_at":"2026-07-28T06:46:30Z","abstract_excerpt":"Generative recommendation formulates recommendation task into an SID sequence autoregressive generation paradigm, but the decoding process is often dominated by generation likelihood. This may conflict with real-world business objectives, where high-value candidates can receive low generation probability and be pruned early during beam search. Existing reranking or training-time alignment methods either intervene too late or require costly model retraining when business preferences change. To this end, we propose \\textbf{R}eward \\textbf{G}uided \\textbf{D}ecoding, named \\textbf{RGD}, a controll"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.25344","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2607.25344/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2607.25344","created_at":"2026-07-29T01:25:10.401979+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.25344v1","created_at":"2026-07-29T01:25:10.401979+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.25344","created_at":"2026-07-29T01:25:10.401979+00:00"},{"alias_kind":"pith_short_12","alias_value":"DC2Q7SJTYY5G","created_at":"2026-07-29T01:25:10.401979+00:00"},{"alias_kind":"pith_short_16","alias_value":"DC2Q7SJTYY5GXHS6","created_at":"2026-07-29T01:25:10.401979+00:00"},{"alias_kind":"pith_short_8","alias_value":"DC2Q7SJT","created_at":"2026-07-29T01:25:10.401979+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2","json":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2.json","graph_json":"https://pith.science/api/pith-number/DC2Q7SJTYY5GXHS6FUHLCUCUP2/graph.json","events_json":"https://pith.science/api/pith-number/DC2Q7SJTYY5GXHS6FUHLCUCUP2/events.json","paper":"https://pith.science/paper/DC2Q7SJT"},"agent_actions":{"view_html":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2","download_json":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2.json","view_paper":"https://pith.science/paper/DC2Q7SJT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.25344&json=true","fetch_graph":"https://pith.science/api/pith-number/DC2Q7SJTYY5GXHS6FUHLCUCUP2/graph.json","fetch_events":"https://pith.science/api/pith-number/DC2Q7SJTYY5GXHS6FUHLCUCUP2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2/action/storage_attestation","attest_author":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2/action/author_attestation","sign_citation":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2/action/citation_signature","submit_replication":"https://pith.science/pith/DC2Q7SJTYY5GXHS6FUHLCUCUP2/action/replication_record"}},"created_at":"2026-07-29T01:25:10.401979+00:00","updated_at":"2026-07-29T01:25:10.401979+00:00"}