{"_canonicalization":{"envelope_id":"axm_ + sha256(envelope minus {signature, axiom_id, anchors})","envelope_signature":"ed25519(envelope minus {signature, axiom_id})","json":"sort_keys=True, separators=(',',':'), ensure_ascii=False, allow_nan=False, utf-8","leaf_hash":"sha256(0x00 || canonical_json(envelope_full))","seal_signature":"ed25519(seal minus {signature, sig_algorithm})"},"axiom_id":"axm_6b8c29949606c12864d5b97f517777cb8311ec97f688b1ebddb82a1603672e1f","bitcoin_anchor":{"bitcoin_attestations":[],"calendar_attestations":[],"ots_url":"","stamped_at":"","status":"pending_next_stamp"},"envelope":{"anchors":[{"chain":"crovia.axiom_graph","height":0,"merkle_proof":"spider_vendor_press_v1","root_at_anchor":"spider_vendor_press_v1"}],"axiom_id":"axm_6b8c29949606c12864d5b97f517777cb8311ec97f688b1ebddb82a1603672e1f","axiom_type":"AX.OBS","body":{"axiom_subtype":"news.vendor_press.v1","category":"news","fingerprint":"98f6a0de22d8c6f8a66250abe51e21e28a478371c8a500f107b8501724a48bbb","published":"Wed, 01 Jul 2026 00:00:00 -0400","receipt_hash":"98f6a0de22d8c6f8a66250abe51e21e28a478371c8a500f107b8501724a48bbb","schema":"spider.news.vendor_press.v1","spider":"vendor_press","spider_record":{"axiom_subtype":"news.vendor_press.v1","category":"news","decision_hint":"POSITIVE","envelope_target":"AX.OBS","fingerprint":"98f6a0de22d8c6f8a66250abe51e21e28a478371c8a500f107b8501724a48bbb","observed_at":"2026-07-01T04:43:38.812993Z","parent_run_hash":"0e10ec7d671a375a4e18d8645653df2b4352c82fe16fae2a400af6f7634d6699","published":"Wed, 01 Jul 2026 00:00:00 -0400","runtime_version":"0.1.0","schema":"spider.news.vendor_press.v1","source_status":200,"source_url":"https://export.arxiv.org/rss/cs.AI","spider":"vendor_press","summary_excerpt":"arXiv:2603.12893v2 Announce Type: replace-cross \nAbstract: Reinforcement learning (RL) has become a standard technique for post-training diffusion-based image synthesis models, as it enables learning from reward signals to explicitly improve desirable aspects such as image quality and prompt alignment. In this paper, we propose an online RL variant that reduces the variance in the model updates by sampling paired trajectories and pulling the flow velocity in the direction of the more favorable image. Unlike existing methods that treat each sampling step as a separate policy action, we consider the entire sampling process as a single action. We experiment with both high-quality vision language models and off-the-shelf quality metrics for rewards, and evaluate the outputs using a broad set of metrics. Our method converges faster and yields higher output quality and prompt alignment than previous approaches.","title":"Finite Difference Flow Optimization for RL Post-Training of Text-to-Image Models","url":"https://arxiv.org/abs/2603.12893","vendor":"arxiv_cs_ai"},"summary":"arXiv:2603.12893v2 Announce Type: replace-cross \nAbstract: Reinforcement learning (RL) has become a standard technique for post-training diffusion-based image synthesis models, as it enables learning from reward signals to explicitly improve desirable aspects such as image quality and prompt alignment. In this paper, we propose an online RL variant that reduces the variance in the model updates by sampling paired trajectories and pulling the flow velocity in the direction of the more favorable image. Unlike existing methods that treat each sampling step as a separate policy action, we consider the entire sampling process as a single action. We experiment with both high-quality vision language models and off-the-shelf quality metrics for rewards, and evaluate the outputs using a broad set of metrics. Our method converges faster and yields higher output quality and prompt alignment than previous approaches.","title":"Finite Difference Flow Optimization for RL Post-Training of Text-to-Image Models","vendor":"arxiv_cs_ai"},"confidence":{"method":"deterministic"},"decision":"POSITIVE","issued_at":"2026-07-01T04:43:38Z","notes":"Spider vendor_press (news) news.vendor_press.v1","object":{"captured_by":"crovia.spider.vendor_press","primary_source_url":"https://arxiv.org/abs/2603.12893"},"predecessors":[],"schema":"crovia.axiom.v1","signature":"ed25519:379b22ca5c40693d93e6e8693cf44f1534224d9de23501711bb2e9888178b0c12824a500e27fdbfc12bd9b43e02bfd058e24a731a1c216e2376423b0f3deae09","signer":"crovia.substrate","subject":{"observed_at":"2026-07-01T04:43:38Z","source_collector":"spider:vendor_press","target_id":"https://arxiv.org/abs/2603.12893"},"tsa":{"authority":"crovia.substrate.bootstrap","rfc3161_token":"{\"kind\":\"crovia.bootstrap.tsa\",\"source_jsonl\":\"/opt/crovia/spider/data/news/vendor_press_v1.jsonl\",\"source_seal_merkle_root\":\"spider_vendor_press_v1\",\"upgrade_path\":\"Sessione H \\u2014 OpenTimestamps weekly anchor\"}"},"zk_mode":"clear","zk_proof":null},"ledger":{"leaf_hash":"8a34d3ccdcaf79e8ee03cd79757225ca57f10ec76302ab527fa17729ff6c7f4d","leaf_index":268705,"ledger_path":"/opt/crovia/substrate/axiom_ledger.jsonl"},"merkle_proof":{"hash_alg":"sha256","leaf_prefix":"0x00","node_prefix":"0x01","odd_leaf_rule":"duplicate_last","path":[{"sibling":"6ef65a79c8d388f6f3f29bedc2cb2ef68f84d9de87c8c07e1dcd7036d56709a9","side":"left"},{"sibling":"c1194ff657d67d6ed0f806a6ad1fc9aa7f76d805b5fb768ffcdfc12573c48a74","side":"right"},{"sibling":"728c9fdea8db5622a17b3be6b25f0b66cb5e378e038593616c8e56da7eff7229","side":"right"},{"sibling":"27bae92ea8c14db860b91404c28c1130101433ed972fbef74b62e2faf6bffe6b","side":"right"},{"sibling":"d9b873370c92e4dc69826859ea55f9515cd72414e290a0e766255ee3ffed8394","side":"right"},{"sibling":"7a1f969ec5b4bf9f076acec001dab3c248a82031e6ffe639dc95059da328f9e9","side":"left"},{"sibling":"4b40722a8676aba99a1cededb6270927be9c50d0f94c30aa826fa53a7a32b4a6","side":"right"},{"sibling":"58c3799598a549d54786b91e80a9c8daea4edcb3c27437b6b16ccadf6dc2889e","side":"left"},{"sibling":"4b14e8ccae46b588748944ffaf410222f66cebc2ac58d936cbe16847b7be9508","side":"left"},{"sibling":"bbd9a20451913e8c7b910f616d5661d9b281a94af70d201ba50b3112d429521b","side":"right"},{"sibling":"85af80e45748c1e0da1e2f42d9d66da8996016888a034f20eb17ff0a73b69eab","side":"right"},{"sibling":"4575fde969d1d9a2984cc01a37ac8441238f74527d42874272dc5582dadebb4f","side":"left"},{"sibling":"f536de281672cbf0b583a3dc46faef1de2823b72bd9265c7b60e889131cc268d","side":"left"},{"sibling":"95b8b0f67237052a17c41fa8cbdce2b79bcb5aeeba3fdb239d4c4497e598115d","side":"right"},{"sibling":"c2f351f771cee329448890504d9436792ba482e50250ca9a19289311131f96c8","side":"right"},{"sibling":"c39bfb2e911ca37ae997690bfc04128ae32e6806ee1cb3908781a7e6685022a0","side":"right"},{"sibling":"a101b4c60ef6854ac3d750eef02d8e2e06c153284b5ecb7111302f97eb129797","side":"right"},{"sibling":"eae2a3de5cb35455ad60125e196cfadaba8a590c53146e95028469f53f70349c","side":"right"},{"sibling":"1cecb7f447febd025aac272837c80de218aecc6485d2395a509b2a1f1b9c746e","side":"left"}]},"schema":"crovia.axiom_proof.v1","seal":{"first_collector_run_id":"","first_receipt_hash":"","jsonl_path":"/opt/crovia/substrate/axiom_ledger.jsonl","key_id":"430895f101d38164","last_collector_run_id":"","last_receipt_hash":"","leaf_count":268860,"merkle_root":"d098f25810d0569730b6c0e170d57f329a70359d483b47874b5f2fc51d23de65","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","run_id":"hourly_json_retrofit_20260701T053701Z","schema":"crovia.seal.v1","seal_family_version":"crovia-seal-family/1","seal_kind":"substrate_batch","sealed_at":"2026-07-01T05:38:05Z","sig_algorithm":"ed25519","signature":"64f37f0e3df0556baa55b924a736cab8005643fa0ab65b503f22407de30eab293e6e0bd62904270fda38d05084bb350c8c0d0aadb2c5a6bae5f1c54598e6d30c","signer_version":"1.1.0"},"trust_root":{"key_id":"430895f101d38164","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","signature_algorithm":"ed25519","url":"/registry/canon/TRUST_ROOT.md"},"verifier":{"spec":"/registry/canon/AXIOM_RECEIPT_v1.md","url":"/v/axm_6b8c29949606c12864d5b97f517777cb8311ec97f688b1ebddb82a1603672e1f"}}