{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:T4R735ZMZIVWNXGW5FM3USCXR5","short_pith_number":"pith:T4R735ZM","schema_version":"1.0","canonical_sha256":"9f23fdf72cca2b66dcd6e959ba48578f57ce50fd9fe354ab79bf52c4288a8af2","source":{"kind":"arxiv","id":"2506.09202","version":2},"attestation_state":"computed","paper":{"title":"Policy-Based Trajectory Clustering in Offline Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hao Hu, Simon Shaolei Du, Xinqi Wang","submitted_at":"2025-06-10T19:44:48Z","abstract_excerpt":"We introduce a novel task of clustering trajectories from offline reinforcement learning (RL) datasets, where each cluster center represents the policy that generated its trajectories. By leveraging the connection between the KL-divergence of offline trajectory distributions and a mixture of policy-induced distributions, we formulate a natural clustering objective. To solve this, we propose Policy-Guided K-means (PG-Kmeans) and Centroid-Attracted Autoencoder (CAAE). PG-Kmeans iteratively trains behavior cloning (BC) policies and assigns trajectories based on policy generation probabilities, wh"},"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":"2506.09202","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-10T19:44:48Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"6782ed248ca8cedcc82aeea21890a60d0794c8cac237159ac18e8fff700ba337","abstract_canon_sha256":"698bf3bfa2e7f7c6c00b4b460361a2bd52fe4efaadae84fcb00ebc205fa39400"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:20:14.531835Z","signature_b64":"6aWBUWenEOeUopzv3RDIYF36SVlvfArUmwTDsui9tMOygeA/s+nQUlq8kiu5hrIL4vyi+1jZKE2wu3xiHP3/Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9f23fdf72cca2b66dcd6e959ba48578f57ce50fd9fe354ab79bf52c4288a8af2","last_reissued_at":"2026-07-05T11:20:14.531286Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:20:14.531286Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Policy-Based Trajectory Clustering in Offline Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hao Hu, Simon Shaolei Du, Xinqi Wang","submitted_at":"2025-06-10T19:44:48Z","abstract_excerpt":"We introduce a novel task of clustering trajectories from offline reinforcement learning (RL) datasets, where each cluster center represents the policy that generated its trajectories. By leveraging the connection between the KL-divergence of offline trajectory distributions and a mixture of policy-induced distributions, we formulate a natural clustering objective. To solve this, we propose Policy-Guided K-means (PG-Kmeans) and Centroid-Attracted Autoencoder (CAAE). PG-Kmeans iteratively trains behavior cloning (BC) policies and assigns trajectories based on policy generation probabilities, wh"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.09202","kind":"arxiv","version":2},"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/2506.09202/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":"2506.09202","created_at":"2026-07-05T11:20:14.531360+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.09202v2","created_at":"2026-07-05T11:20:14.531360+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.09202","created_at":"2026-07-05T11:20:14.531360+00:00"},{"alias_kind":"pith_short_12","alias_value":"T4R735ZMZIVW","created_at":"2026-07-05T11:20:14.531360+00:00"},{"alias_kind":"pith_short_16","alias_value":"T4R735ZMZIVWNXGW","created_at":"2026-07-05T11:20:14.531360+00:00"},{"alias_kind":"pith_short_8","alias_value":"T4R735ZM","created_at":"2026-07-05T11:20:14.531360+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24601","citing_title":"ASALT: Adaptive State Alignment for Lateral Transfer in Multi-agent Reinforcement Learning","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12200","citing_title":"Implicit Neural Representations of Individual Behavior","ref_index":92,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5","json":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5.json","graph_json":"https://pith.science/api/pith-number/T4R735ZMZIVWNXGW5FM3USCXR5/graph.json","events_json":"https://pith.science/api/pith-number/T4R735ZMZIVWNXGW5FM3USCXR5/events.json","paper":"https://pith.science/paper/T4R735ZM"},"agent_actions":{"view_html":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5","download_json":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5.json","view_paper":"https://pith.science/paper/T4R735ZM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.09202&json=true","fetch_graph":"https://pith.science/api/pith-number/T4R735ZMZIVWNXGW5FM3USCXR5/graph.json","fetch_events":"https://pith.science/api/pith-number/T4R735ZMZIVWNXGW5FM3USCXR5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5/action/storage_attestation","attest_author":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5/action/author_attestation","sign_citation":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5/action/citation_signature","submit_replication":"https://pith.science/pith/T4R735ZMZIVWNXGW5FM3USCXR5/action/replication_record"}},"created_at":"2026-07-05T11:20:14.531360+00:00","updated_at":"2026-07-05T11:20:14.531360+00:00"}