{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6XQQV3ODNGDQFHKNRNJIP4ZCE7","short_pith_number":"pith:6XQQV3OD","schema_version":"1.0","canonical_sha256":"f5e10aedc36987029d4d8b5287f32227edadbcc2332e90b4b1734fc0f595b83f","source":{"kind":"arxiv","id":"2605.27736","version":1},"attestation_state":"computed","paper":{"title":"Explicit Critic Guidance for Aligning Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Ceyuan Yang, Qihang Zhang, Zhengyang Liang","submitted_at":"2026-05-26T22:20:51Z","abstract_excerpt":"Online reinforcement learning is becoming increasingly important for aligning diffusion models with non-differentiable objectives. However, existing methods still face limitations in assigning fine-grained credit along denoising trajectories and in realizing stable value-based optimization. We propose a state-aligned latent actor-critic framework for diffusion post-training, in which the diffusion model serves as its own timestep-conditioned value function and predicts values directly on noisy latent states. This enables trajectory-level PPO training, supports stable actor-critic optimization "},"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":"2605.27736","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-26T22:20:51Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"6a6995e87f850c181a2572167533fdb892d9f67b9aa3cbf306bc10e44eeee489","abstract_canon_sha256":"4f4ba94cc16ee86bc3aded616e56e26aee11011719517cd34eb4180417ef9371"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:47.565079Z","signature_b64":"UojY2qIo5WBYJ/KrwsiV+DhtFCckDCVr0874caIbsdCzYZ6TZiRIH1oNEnDg9w+mmIh48LXp0kAq9Fu0vNJOBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f5e10aedc36987029d4d8b5287f32227edadbcc2332e90b4b1734fc0f595b83f","last_reissued_at":"2026-05-28T01:04:47.564706Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:47.564706Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Explicit Critic Guidance for Aligning Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Ceyuan Yang, Qihang Zhang, Zhengyang Liang","submitted_at":"2026-05-26T22:20:51Z","abstract_excerpt":"Online reinforcement learning is becoming increasingly important for aligning diffusion models with non-differentiable objectives. However, existing methods still face limitations in assigning fine-grained credit along denoising trajectories and in realizing stable value-based optimization. We propose a state-aligned latent actor-critic framework for diffusion post-training, in which the diffusion model serves as its own timestep-conditioned value function and predicts values directly on noisy latent states. This enables trajectory-level PPO training, supports stable actor-critic optimization "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.27736","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/2605.27736/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":"2605.27736","created_at":"2026-05-28T01:04:47.564760+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.27736v1","created_at":"2026-05-28T01:04:47.564760+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.27736","created_at":"2026-05-28T01:04:47.564760+00:00"},{"alias_kind":"pith_short_12","alias_value":"6XQQV3ODNGDQ","created_at":"2026-05-28T01:04:47.564760+00:00"},{"alias_kind":"pith_short_16","alias_value":"6XQQV3ODNGDQFHKN","created_at":"2026-05-28T01:04:47.564760+00:00"},{"alias_kind":"pith_short_8","alias_value":"6XQQV3OD","created_at":"2026-05-28T01:04:47.564760+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/6XQQV3ODNGDQFHKNRNJIP4ZCE7","json":"https://pith.science/pith/6XQQV3ODNGDQFHKNRNJIP4ZCE7.json","graph_json":"https://pith.science/api/pith-number/6XQQV3ODNGDQFHKNRNJIP4ZCE7/graph.json","events_json":"https://pith.science/api/pith-number/6XQQV3ODNGDQFHKNRNJIP4ZCE7/events.json","paper":"https://pith.science/paper/6XQQV3OD"},"agent_actions":{"view_html":"https://pith.science/pith/6XQQV3ODNGDQFHKNRNJIP4ZCE7","download_json":"https://pith.science/pith/6XQQV3ODNGDQFHKNRNJIP4ZCE7.json","view_paper":"https://pith.science/paper/6XQQV3OD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.27736&json=true","fetch_graph":"https://pith.science/api/pith-number/6XQQV3ODNGDQFHKNRNJIP4ZCE7/graph.json","fetch_events":"https://pith.science/api/pith-number/6XQQV3ODNGDQFHKNRNJIP4ZCE7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6XQQV3ODNGDQFHKNRNJIP4ZCE7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6XQQV3ODNGDQFHKNRNJIP4ZCE7/action/storage_attestation","attest_author":"https://pith.science/pith/6XQQV3ODNGDQFHKNRNJIP4ZCE7/action/author_attestation","sign_citation":"https://pith.science/pith/6XQQV3ODNGDQFHKNRNJIP4ZCE7/action/citation_signature","submit_replication":"https://pith.science/pith/6XQQV3ODNGDQFHKNRNJIP4ZCE7/action/replication_record"}},"created_at":"2026-05-28T01:04:47.564760+00:00","updated_at":"2026-05-28T01:04:47.564760+00:00"}