{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:P4HUNZARFD5CNASISNWE63AYQN","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":"ae46796841d76b3575eba32104b569b38dee8c7c75a0de345a25ac3d84abef1a","cross_cats_sorted":["cs.RO"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-06T09:03:41Z","title_canon_sha256":"f94ebc986218d4761adc4446fc388da88b050823e912340b7a2eef21558e682f"},"schema_version":"1.0","source":{"id":"2604.04539","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.04539","created_at":"2026-05-20T00:00:37Z"},{"alias_kind":"arxiv_version","alias_value":"2604.04539v2","created_at":"2026-05-20T00:00:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.04539","created_at":"2026-05-20T00:00:37Z"},{"alias_kind":"pith_short_12","alias_value":"P4HUNZARFD5C","created_at":"2026-05-20T00:00:37Z"},{"alias_kind":"pith_short_16","alias_value":"P4HUNZARFD5CNASI","created_at":"2026-05-20T00:00:37Z"},{"alias_kind":"pith_short_8","alias_value":"P4HUNZAR","created_at":"2026-05-20T00:00:37Z"}],"graph_snapshots":[{"event_id":"sha256:299aa5795f3c97ff3f745fa60582e5d9baf9102100941e011d91ae89d7954b34","target":"graph","created_at":"2026-05-20T00:00:37Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Across over 60 tasks in 10 simulators, FlashSAC consistently outperforms PPO and strong off-policy baselines in both final performance and training efficiency, with the largest gains on high-dimensional tasks such as dexterous manipulation. In sim-to-real humanoid locomotion, FlashSAC reduces training time from hours to minutes."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That explicitly bounding weight, feature, and gradient norms will sufficiently curb critic error accumulation in high-dimensional spaces without removing the capacity needed for accurate value estimation or policy improvement."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"FlashSAC improves training speed and final performance of off-policy RL on high-dimensional robot tasks by reducing update frequency, increasing model scale, and bounding norms to limit critic error accumulation."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"FlashSAC stabilizes off-policy RL for high-dimensional robot control by cutting gradient updates and bounding norms to limit critic errors."}],"snapshot_sha256":"350de7fc0cf5087e99812e76d522f097aa206c6fc5bd272d14c2d18301905f75"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"585797ed27890a08802b5db6540d6dc87e1256c228071a6ff52fcf14ba011a2f"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.04539/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement learning (RL) is a core approach for robot control when expert demonstrations are unavailable. On-policy methods such as Proximal Policy Optimization (PPO) are widely used for their stability, but their reliance on narrowly distributed on-policy data limits accurate policy evaluation in high-dimensional state and action spaces. Off-policy methods can overcome this limitation by learning from a broader state-action distribution, yet suffer from slow convergence and instability, as fitting a value function over diverse data requires many gradient updates, causing critic errors to a","authors_text":"Danica Kragic, Daniel Palenicek, Donghu Kim, Florian Vogt, Hojoon Lee, I Made Aswin Nahendra, Jaegul Choo, Jan Peters, Kinam Kim, Minho Park, Sehee Min, Takuma Seno, Youngdo Lee","cross_cats":["cs.RO"],"headline":"FlashSAC stabilizes off-policy RL for high-dimensional robot control by cutting gradient updates and bounding norms to limit critic errors.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-06T09:03:41Z","title":"FlashSAC: Fast and Stable Off-Policy Reinforcement Learning for High-Dimensional Robot Control"},"references":{"count":94,"internal_anchors":18,"resolved_work":94,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Loss of plasticity in continual deep reinforcement learning","work_id":"7ecff814-af37-405d-a70f-9c1fb4554063","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Learning dexterous in-hand manipulation.The International Journal of Robotics Research, 39(1):3–20, 2020","work_id":"d1093acb-7a4b-42e8-aa29-51fcfd4bab71","year":2020},{"cited_arxiv_id":"1708.05866","doi":"","is_internal_anchor":true,"ref_index":3,"title":"A brief survey of deep reinforcement learning","work_id":"3b2842ca-9c8f-4b1a-a615-22c29e63726b","year":2017},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Genesis: A generative and universal physics engine for robotics and beyond, December 2024","work_id":"cb89394b-6f98-44ca-8234-f0020e9373e8","year":2024},{"cited_arxiv_id":"1607.06450","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Layer Normalization","work_id":"20a2d720-0046-4c7c-bcd6-327ec8143f69","year":2016}],"snapshot_sha256":"9ea9fa4581b88687296c237e979740aa6eb68ac1e53414cdc29123fcc43c1529"},"source":{"id":"2604.04539","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-19T17:04:49.619467Z","id":"4f1dfd44-0e81-404c-95aa-a813e0e143f9","model_set":{"reader":"grok-4.3"},"one_line_summary":"FlashSAC improves training speed and final performance of off-policy RL on high-dimensional robot tasks by reducing update frequency, increasing model scale, and bounding norms to limit critic error accumulation.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"FlashSAC stabilizes off-policy RL for high-dimensional robot control by cutting gradient updates and bounding norms to limit critic errors.","strongest_claim":"Across over 60 tasks in 10 simulators, FlashSAC consistently outperforms PPO and strong off-policy baselines in both final performance and training efficiency, with the largest gains on high-dimensional tasks such as dexterous manipulation. In sim-to-real humanoid locomotion, FlashSAC reduces training time from hours to minutes.","weakest_assumption":"That explicitly bounding weight, feature, and gradient norms will sufficiently curb critic error accumulation in high-dimensional spaces without removing the capacity needed for accurate value estimation or policy improvement."}},"verdict_id":"4f1dfd44-0e81-404c-95aa-a813e0e143f9"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:9e031a3f31ea7fff04ac2c683de5c3dc81c1550920bbba01060fcc30b46bf936","target":"record","created_at":"2026-05-20T00:00:37Z","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":"ae46796841d76b3575eba32104b569b38dee8c7c75a0de345a25ac3d84abef1a","cross_cats_sorted":["cs.RO"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-06T09:03:41Z","title_canon_sha256":"f94ebc986218d4761adc4446fc388da88b050823e912340b7a2eef21558e682f"},"schema_version":"1.0","source":{"id":"2604.04539","kind":"arxiv","version":2}},"canonical_sha256":"7f0f46e41128fa268248936c4f6c188351bd41ef445d3cfd1aa858d1d78ab95f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7f0f46e41128fa268248936c4f6c188351bd41ef445d3cfd1aa858d1d78ab95f","first_computed_at":"2026-05-20T00:00:37.674180Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:00:37.674180Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"xmugFJTdgXJcb6ywZpjHgbuxSLPDlP/N/vTCwmAg8wTkQpH9NrTfxUZhzpjR8XncroobcUGLRITv3mazzoetBQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:00:37.674882Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.04539","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9e031a3f31ea7fff04ac2c683de5c3dc81c1550920bbba01060fcc30b46bf936","sha256:299aa5795f3c97ff3f745fa60582e5d9baf9102100941e011d91ae89d7954b34"],"state_sha256":"c780d1f5f1fd5099bcccd90fa81ba1f358674e8878cc6e62ce7a6bc8ac9a968b"}