{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SI3M65ELF4MCZ5KA43M7O3TION","short_pith_number":"pith:SI3M65EL","schema_version":"1.0","canonical_sha256":"9236cf748b2f182cf540e6d9f76e687348792d370ebbfd30d7cad8cc8685d51f","source":{"kind":"arxiv","id":"2606.13795","version":2},"attestation_state":"computed","paper":{"title":"DiPOD: Diffusion Policy Optimization without Drifting Apart","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Angjoo Kanazawa, Haiwen Feng, Haozhe Jiang, Jiantao Jiao, Nika Haghtalab, Pieter Abbeel","submitted_at":"2026-06-11T18:06:04Z","abstract_excerpt":"RL post-training has become increasingly pivotal for improving diffusion policies, but existing diffusion policy-gradient methods are often unstable and cannot achieve reliable policy improvement. We identify the cause as the double-drift phenomenon: optimizing a variational surrogate can let the ELBO separate from the true log-likelihood, which then makes the resulting proxy policy gradient misaligned with the true policy gradient of expected return. We propose \\textbf{DiPOD}, a diffusion policy optimization framework that maintains tight-bound behavior throughout training by interleaving sel"},"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":"2606.13795","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-11T18:06:04Z","cross_cats_sorted":[],"title_canon_sha256":"5cc51eca7b61b70fecff8a8d1ec97624ad1ba8faab38975ee36d993751af1a71","abstract_canon_sha256":"ca8df28d4a73c6214a275a32f4be489428ee5f87e2a3023562a3c88c6c45b443"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:12:06.521540Z","signature_b64":"flTGEjq+fMaUQFoZY+xwz4t61XIEZWaVgl2GDkb6L0GxAkFhYzzmtyw6bCdHBzp34pzhXmzuoQBvA4CsDqdbAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9236cf748b2f182cf540e6d9f76e687348792d370ebbfd30d7cad8cc8685d51f","last_reissued_at":"2026-06-19T16:12:06.521130Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:12:06.521130Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DiPOD: Diffusion Policy Optimization without Drifting Apart","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Angjoo Kanazawa, Haiwen Feng, Haozhe Jiang, Jiantao Jiao, Nika Haghtalab, Pieter Abbeel","submitted_at":"2026-06-11T18:06:04Z","abstract_excerpt":"RL post-training has become increasingly pivotal for improving diffusion policies, but existing diffusion policy-gradient methods are often unstable and cannot achieve reliable policy improvement. We identify the cause as the double-drift phenomenon: optimizing a variational surrogate can let the ELBO separate from the true log-likelihood, which then makes the resulting proxy policy gradient misaligned with the true policy gradient of expected return. We propose \\textbf{DiPOD}, a diffusion policy optimization framework that maintains tight-bound behavior throughout training by interleaving sel"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.13795","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/2606.13795/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":"2606.13795","created_at":"2026-06-19T16:12:06.521180+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.13795v2","created_at":"2026-06-19T16:12:06.521180+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.13795","created_at":"2026-06-19T16:12:06.521180+00:00"},{"alias_kind":"pith_short_12","alias_value":"SI3M65ELF4MC","created_at":"2026-06-19T16:12:06.521180+00:00"},{"alias_kind":"pith_short_16","alias_value":"SI3M65ELF4MCZ5KA","created_at":"2026-06-19T16:12:06.521180+00:00"},{"alias_kind":"pith_short_8","alias_value":"SI3M65EL","created_at":"2026-06-19T16:12:06.521180+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/SI3M65ELF4MCZ5KA43M7O3TION","json":"https://pith.science/pith/SI3M65ELF4MCZ5KA43M7O3TION.json","graph_json":"https://pith.science/api/pith-number/SI3M65ELF4MCZ5KA43M7O3TION/graph.json","events_json":"https://pith.science/api/pith-number/SI3M65ELF4MCZ5KA43M7O3TION/events.json","paper":"https://pith.science/paper/SI3M65EL"},"agent_actions":{"view_html":"https://pith.science/pith/SI3M65ELF4MCZ5KA43M7O3TION","download_json":"https://pith.science/pith/SI3M65ELF4MCZ5KA43M7O3TION.json","view_paper":"https://pith.science/paper/SI3M65EL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.13795&json=true","fetch_graph":"https://pith.science/api/pith-number/SI3M65ELF4MCZ5KA43M7O3TION/graph.json","fetch_events":"https://pith.science/api/pith-number/SI3M65ELF4MCZ5KA43M7O3TION/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SI3M65ELF4MCZ5KA43M7O3TION/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SI3M65ELF4MCZ5KA43M7O3TION/action/storage_attestation","attest_author":"https://pith.science/pith/SI3M65ELF4MCZ5KA43M7O3TION/action/author_attestation","sign_citation":"https://pith.science/pith/SI3M65ELF4MCZ5KA43M7O3TION/action/citation_signature","submit_replication":"https://pith.science/pith/SI3M65ELF4MCZ5KA43M7O3TION/action/replication_record"}},"created_at":"2026-06-19T16:12:06.521180+00:00","updated_at":"2026-06-19T16:12:06.521180+00:00"}