{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7TGMMVSJLI3FOOAINZGZYOCQDC","short_pith_number":"pith:7TGMMVSJ","schema_version":"1.0","canonical_sha256":"fcccc656495a365738086e4d9c385018a37da59b24081c23c6122c486fb2eeb9","source":{"kind":"arxiv","id":"2608.02951","version":1},"attestation_state":"computed","paper":{"title":"SP3O: Reinforcement Learning from Segment Preferences without Reward Modeling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Evan Assmus, Lei Ying, Qining Zhang","submitted_at":"2026-08-03T23:28:06Z","abstract_excerpt":"Preference-based reinforcement learning (PbRL) for general stochastic MDPs often requires training a reward model. Existing reward-model-free methods are either restricted to bandits or deterministic MDPs, such as DPO or P3O, or use zeroth-order, gradient-free optimization, which in general exhibits a slower convergence rate than gradient-based algorithms. Furthermore, existing reward-model-free preference-based RL algorithms almost exclusively use trajectory-level feedback, which can require significant effort from a human evaluator when trajectories are long. On the other hand, segments are "},"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":"2608.02951","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-03T23:28:06Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"04829bceb7f7f453107e38e70a01e31ed3bc7dc7a6fc64423a5da441447b587c","abstract_canon_sha256":"3198d058a4fe3c1807f6a08cea0a10962534c102df7d0631dc5dc85762d69108"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-05T00:44:13.262432Z","signature_b64":"SMo74aNqwasFjxDhfkmFRYiW1g5cOrD9ctLOTyTE9qxe2pmy12VKcR0nPMt3LQN/XTjQgAp/8LvOI8xQvrvmDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fcccc656495a365738086e4d9c385018a37da59b24081c23c6122c486fb2eeb9","last_reissued_at":"2026-08-05T00:44:13.260063Z","signature_status":"signed_v1","first_computed_at":"2026-08-05T00:44:13.260063Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SP3O: Reinforcement Learning from Segment Preferences without Reward Modeling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Evan Assmus, Lei Ying, Qining Zhang","submitted_at":"2026-08-03T23:28:06Z","abstract_excerpt":"Preference-based reinforcement learning (PbRL) for general stochastic MDPs often requires training a reward model. Existing reward-model-free methods are either restricted to bandits or deterministic MDPs, such as DPO or P3O, or use zeroth-order, gradient-free optimization, which in general exhibits a slower convergence rate than gradient-based algorithms. Furthermore, existing reward-model-free preference-based RL algorithms almost exclusively use trajectory-level feedback, which can require significant effort from a human evaluator when trajectories are long. On the other hand, segments are "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.02951","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/2608.02951/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":"2608.02951","created_at":"2026-08-05T00:44:13.260854+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.02951v1","created_at":"2026-08-05T00:44:13.260854+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.02951","created_at":"2026-08-05T00:44:13.260854+00:00"},{"alias_kind":"pith_short_12","alias_value":"7TGMMVSJLI3F","created_at":"2026-08-05T00:44:13.260854+00:00"},{"alias_kind":"pith_short_16","alias_value":"7TGMMVSJLI3FOOAI","created_at":"2026-08-05T00:44:13.260854+00:00"},{"alias_kind":"pith_short_8","alias_value":"7TGMMVSJ","created_at":"2026-08-05T00:44:13.260854+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/7TGMMVSJLI3FOOAINZGZYOCQDC","json":"https://pith.science/pith/7TGMMVSJLI3FOOAINZGZYOCQDC.json","graph_json":"https://pith.science/api/pith-number/7TGMMVSJLI3FOOAINZGZYOCQDC/graph.json","events_json":"https://pith.science/api/pith-number/7TGMMVSJLI3FOOAINZGZYOCQDC/events.json","paper":"https://pith.science/paper/7TGMMVSJ"},"agent_actions":{"view_html":"https://pith.science/pith/7TGMMVSJLI3FOOAINZGZYOCQDC","download_json":"https://pith.science/pith/7TGMMVSJLI3FOOAINZGZYOCQDC.json","view_paper":"https://pith.science/paper/7TGMMVSJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.02951&json=true","fetch_graph":"https://pith.science/api/pith-number/7TGMMVSJLI3FOOAINZGZYOCQDC/graph.json","fetch_events":"https://pith.science/api/pith-number/7TGMMVSJLI3FOOAINZGZYOCQDC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7TGMMVSJLI3FOOAINZGZYOCQDC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7TGMMVSJLI3FOOAINZGZYOCQDC/action/storage_attestation","attest_author":"https://pith.science/pith/7TGMMVSJLI3FOOAINZGZYOCQDC/action/author_attestation","sign_citation":"https://pith.science/pith/7TGMMVSJLI3FOOAINZGZYOCQDC/action/citation_signature","submit_replication":"https://pith.science/pith/7TGMMVSJLI3FOOAINZGZYOCQDC/action/replication_record"}},"created_at":"2026-08-05T00:44:13.260854+00:00","updated_at":"2026-08-05T00:44:13.260854+00:00"}