{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GA3KHFDVH3EOIQAVX2MHKM627T","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"24ef5696f1103e19d8de1fbd5aa567d49dc4a04bf4201416f90c07223141e395","cross_cats_sorted":["cs.AI","cs.CV"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T16:32:14Z","title_canon_sha256":"79c00d175876ebf62899fb383bd168913498fd290e7dc9daea278a01c71ac481"},"schema_version":"1.0","source":{"id":"2605.26013","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.26013","created_at":"2026-05-26T02:05:23Z"},{"alias_kind":"arxiv_version","alias_value":"2605.26013v1","created_at":"2026-05-26T02:05:23Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.26013","created_at":"2026-05-26T02:05:23Z"},{"alias_kind":"pith_short_12","alias_value":"GA3KHFDVH3EO","created_at":"2026-05-26T02:05:23Z"},{"alias_kind":"pith_short_16","alias_value":"GA3KHFDVH3EOIQAV","created_at":"2026-05-26T02:05:23Z"},{"alias_kind":"pith_short_8","alias_value":"GA3KHFDV","created_at":"2026-05-26T02:05:23Z"}],"graph_snapshots":[{"event_id":"sha256:f9cc1866fb11d889bd317a73e5a0e54faf31a00e3998196051e51b4af169bb95","target":"graph","created_at":"2026-05-26T02:05:23Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2605.26013/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We introduce AdvantageFlow, a forward-process reinforcement learning algorithm for rectified flow models. Unlike Flow-GRPO, which optimizes the reverse process, we optimize an advantage-weighted forward-process prediction loss. This optimization problem is unstable when advantages are negative and the loss becomes non-convex. We stabilize it by rollout policy regularization, which reduces variance and arises from fitting a local reward-improving target distribution. We evaluate AdvantageFlow on image generation tasks with Stable Diffusion 3.5 Medium. It outperforms both Flow-GRPO and a state-o","authors_text":"Anup Rao, Branislav Kveton, Krishna Kumar Singh, Subhojyoti Mukherjee, Viet Dac Lai","cross_cats":["cs.AI","cs.CV"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T16:32:14Z","title":"AdvantageFlow: Advantage-Weighted Least Squares for RL in Flow Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.26013","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:85a4974fa33b62aa20c950effe2d26ab955f5d3283f77332573325ddb7fb3a76","target":"record","created_at":"2026-05-26T02:05:23Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"24ef5696f1103e19d8de1fbd5aa567d49dc4a04bf4201416f90c07223141e395","cross_cats_sorted":["cs.AI","cs.CV"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T16:32:14Z","title_canon_sha256":"79c00d175876ebf62899fb383bd168913498fd290e7dc9daea278a01c71ac481"},"schema_version":"1.0","source":{"id":"2605.26013","kind":"arxiv","version":1}},"canonical_sha256":"3036a394753ec8e44015be987533dafcde8eafa8253a83fab5258f0d62d3307c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3036a394753ec8e44015be987533dafcde8eafa8253a83fab5258f0d62d3307c","first_computed_at":"2026-05-26T02:05:23.266900Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T02:05:23.266900Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"2yv4u7xcTepPZAG3aCYIRTTj1WWT2jCcS945/VRWNzVva/de3+nOaMVBjmbzaxXOEDChKjW2DB4g3ZzBgx2ECg==","signature_status":"signed_v1","signed_at":"2026-05-26T02:05:23.267626Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.26013","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:85a4974fa33b62aa20c950effe2d26ab955f5d3283f77332573325ddb7fb3a76","sha256:f9cc1866fb11d889bd317a73e5a0e54faf31a00e3998196051e51b4af169bb95"],"state_sha256":"c0152a66465e2f07b4bbfcd35cf11b187acf1c56458abc27a5014b38e6f180f3"}