{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:O65NSA5TMVGGLUXZ364C57JXTZ","short_pith_number":"pith:O65NSA5T","canonical_record":{"source":{"id":"2604.17787","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-04-20T04:25:24Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7ad7e8a3d3f2c48b317c9d28cc1c8737438b7f635cb5beb9d5960ca96ef77e5a","abstract_canon_sha256":"17a023e370d94ad2a9fd2c3abea104c4ad14f980fa2e28df751f7b4ddc586185"},"schema_version":"1.0"},"canonical_sha256":"77bad903b3654c65d2f9dfb82efd379e5668f9ce177ed59701bc5c6fc077b86b","source":{"kind":"arxiv","id":"2604.17787","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.17787","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"arxiv_version","alias_value":"2604.17787v2","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.17787","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"pith_short_12","alias_value":"O65NSA5TMVGG","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"pith_short_16","alias_value":"O65NSA5TMVGGLUXZ","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"pith_short_8","alias_value":"O65NSA5T","created_at":"2026-07-22T01:22:41Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:O65NSA5TMVGGLUXZ364C57JXTZ","target":"record","payload":{"canonical_record":{"source":{"id":"2604.17787","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-04-20T04:25:24Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7ad7e8a3d3f2c48b317c9d28cc1c8737438b7f635cb5beb9d5960ca96ef77e5a","abstract_canon_sha256":"17a023e370d94ad2a9fd2c3abea104c4ad14f980fa2e28df751f7b4ddc586185"},"schema_version":"1.0"},"canonical_sha256":"77bad903b3654c65d2f9dfb82efd379e5668f9ce177ed59701bc5c6fc077b86b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-22T01:22:41.374312Z","signature_b64":"PWRRHUCZJCCBYXraQ/kw1Rfd0CxSocj+y1vOLzg7+5Nc9lYQv8nuLRxIOeIM7hZppoNFeKzuFAKUoAB2eetjAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"77bad903b3654c65d2f9dfb82efd379e5668f9ce177ed59701bc5c6fc077b86b","last_reissued_at":"2026-07-22T01:22:41.373337Z","signature_status":"signed_v1","first_computed_at":"2026-07-22T01:22:41.373337Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.17787","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-22T01:22:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rEwJBaU70Tt+6YUzeHeid6IX9jM36zND2+6jssbzTDXrFjhYAu0moM3YlhtxkFqG4IMeRoUxK8zlgjY0WC4cAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T08:24:34.224492Z"},"content_sha256":"c688d5c98d10ef2a07db5216dfa218525feb343b0abd7f927c154c5fd949de00","schema_version":"1.0","event_id":"sha256:c688d5c98d10ef2a07db5216dfa218525feb343b0abd7f927c154c5fd949de00"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:O65NSA5TMVGGLUXZ364C57JXTZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AnchorRefine: Synergy-Manipulation Based on Trajectory Anchor and Residual Refinement for Vision-Language-Action Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"AnchorRefine factorizes VLA action prediction into a coarse trajectory anchor plus residual refinement to raise manipulation precision.","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Baocai Yin, Chunmian Lin, Daxin Tian, Guixian Qu, Jiapu Wang, Kan Guo, Lanping Qian, Tingzheng Jia, Yongli Hu","submitted_at":"2026-04-20T04:25:24Z","abstract_excerpt":"Precision-critical manipulation requires both global trajectory organization and local execution correction, yet most vision-language-action (VLA) policies generate actions within a single unified space. This monolithic formulation forces macro-level transport and micro-level refinement to be optimized under the same objective, causing large motions to dominate learning while suppressing small but failure-critical corrective signals. In contrast, human manipulation is structured by global movement planning together with continuous local adjustment during execution. Motivated by this principle,"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on LIBERO, CALVIN, and real-robot tasks demonstrate that AnchorRefine consistently improves both regression-based and diffusion-based VLA backbones, yielding gains of up to 7.8% in simulation success rate and 18% in real-world success rate.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that monolithic unified action spaces cause large motions to dominate learning and suppress small corrective signals, and that explicitly factorizing into anchor plus residual will reliably improve geometric and contact precision without new failure modes.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"AnchorRefine factorizes VLA action generation into a trajectory anchor for coarse planning and residual refinement for local corrections, improving success rates by up to 7.8% in simulation and 18% on real robots across LIBERO, CALVIN, and physical tasks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"AnchorRefine factorizes VLA action prediction into a coarse trajectory anchor plus residual refinement to raise manipulation precision.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2cfea222c2b59c0b4f194e25bfbd4fa25e86e1dea9337b24a5f6f48ca827cd74"},"source":{"id":"2604.17787","kind":"arxiv","version":2},"verdict":{"id":"588f7b61-02b8-468f-a41d-d522fbf37258","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T05:07:41.417241Z","strongest_claim":"Experiments on LIBERO, CALVIN, and real-robot tasks demonstrate that AnchorRefine consistently improves both regression-based and diffusion-based VLA backbones, yielding gains of up to 7.8% in simulation success rate and 18% in real-world success rate.","one_line_summary":"AnchorRefine factorizes VLA action generation into a trajectory anchor for coarse planning and residual refinement for local corrections, improving success rates by up to 7.8% in simulation and 18% on real robots across LIBERO, CALVIN, and physical tasks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that monolithic unified action spaces cause large motions to dominate learning and suppress small corrective signals, and that explicitly factorizing into anchor plus residual will reliably improve geometric and contact precision without new failure modes.","pith_extraction_headline":"AnchorRefine factorizes VLA action prediction into a coarse trajectory anchor plus residual refinement to raise manipulation precision."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.17787/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"},"verdict_id":"588f7b61-02b8-468f-a41d-d522fbf37258"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-22T01:22:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VazzuO80UWiNn9AzRQ6pXOZ2u/a5YsRlONv78R6DyAlu5n9dYFAg6lZ2v/+QlpzSHeribACN67JH12KzJ/gHDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T08:24:34.225523Z"},"content_sha256":"cc042f1f98e13a601ca35132b5f2dd319f87cb4af8c3b90c09cf64c9109884ca","schema_version":"1.0","event_id":"sha256:cc042f1f98e13a601ca35132b5f2dd319f87cb4af8c3b90c09cf64c9109884ca"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/O65NSA5TMVGGLUXZ364C57JXTZ/bundle.json","state_url":"https://pith.science/pith/O65NSA5TMVGGLUXZ364C57JXTZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/O65NSA5TMVGGLUXZ364C57JXTZ/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-04T08:24:34Z","links":{"resolver":"https://pith.science/pith/O65NSA5TMVGGLUXZ364C57JXTZ","bundle":"https://pith.science/pith/O65NSA5TMVGGLUXZ364C57JXTZ/bundle.json","state":"https://pith.science/pith/O65NSA5TMVGGLUXZ364C57JXTZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/O65NSA5TMVGGLUXZ364C57JXTZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:O65NSA5TMVGGLUXZ364C57JXTZ","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":"17a023e370d94ad2a9fd2c3abea104c4ad14f980fa2e28df751f7b4ddc586185","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-04-20T04:25:24Z","title_canon_sha256":"7ad7e8a3d3f2c48b317c9d28cc1c8737438b7f635cb5beb9d5960ca96ef77e5a"},"schema_version":"1.0","source":{"id":"2604.17787","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.17787","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"arxiv_version","alias_value":"2604.17787v2","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.17787","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"pith_short_12","alias_value":"O65NSA5TMVGG","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"pith_short_16","alias_value":"O65NSA5TMVGGLUXZ","created_at":"2026-07-22T01:22:41Z"},{"alias_kind":"pith_short_8","alias_value":"O65NSA5T","created_at":"2026-07-22T01:22:41Z"}],"graph_snapshots":[{"event_id":"sha256:cc042f1f98e13a601ca35132b5f2dd319f87cb4af8c3b90c09cf64c9109884ca","target":"graph","created_at":"2026-07-22T01:22:41Z","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":"Experiments on LIBERO, CALVIN, and real-robot tasks demonstrate that AnchorRefine consistently improves both regression-based and diffusion-based VLA backbones, yielding gains of up to 7.8% in simulation success rate and 18% in real-world success rate."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The assumption that monolithic unified action spaces cause large motions to dominate learning and suppress small corrective signals, and that explicitly factorizing into anchor plus residual will reliably improve geometric and contact precision without new failure modes."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"AnchorRefine factorizes VLA action generation into a trajectory anchor for coarse planning and residual refinement for local corrections, improving success rates by up to 7.8% in simulation and 18% on real robots across LIBERO, CALVIN, and physical tasks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"AnchorRefine factorizes VLA action prediction into a coarse trajectory anchor plus residual refinement to raise manipulation precision."}],"snapshot_sha256":"2cfea222c2b59c0b4f194e25bfbd4fa25e86e1dea9337b24a5f6f48ca827cd74"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.17787/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Precision-critical manipulation requires both global trajectory organization and local execution correction, yet most vision-language-action (VLA) policies generate actions within a single unified space. This monolithic formulation forces macro-level transport and micro-level refinement to be optimized under the same objective, causing large motions to dominate learning while suppressing small but failure-critical corrective signals. In contrast, human manipulation is structured by global movement planning together with continuous local adjustment during execution. Motivated by this principle,","authors_text":"Baocai Yin, Chunmian Lin, Daxin Tian, Guixian Qu, Jiapu Wang, Kan Guo, Lanping Qian, Tingzheng Jia, Yongli Hu","cross_cats":["cs.AI"],"headline":"AnchorRefine factorizes VLA action prediction into a coarse trajectory anchor plus residual refinement to raise manipulation precision.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-04-20T04:25:24Z","title":"AnchorRefine: Synergy-Manipulation Based on Trajectory Anchor and Residual Refinement for Vision-Language-Action Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.17787","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-10T05:07:41.417241Z","id":"588f7b61-02b8-468f-a41d-d522fbf37258","model_set":{"reader":"grok-4.3"},"one_line_summary":"AnchorRefine factorizes VLA action generation into a trajectory anchor for coarse planning and residual refinement for local corrections, improving success rates by up to 7.8% in simulation and 18% on real robots across LIBERO, CALVIN, and physical tasks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"AnchorRefine factorizes VLA action prediction into a coarse trajectory anchor plus residual refinement to raise manipulation precision.","strongest_claim":"Experiments on LIBERO, CALVIN, and real-robot tasks demonstrate that AnchorRefine consistently improves both regression-based and diffusion-based VLA backbones, yielding gains of up to 7.8% in simulation success rate and 18% in real-world success rate.","weakest_assumption":"The assumption that monolithic unified action spaces cause large motions to dominate learning and suppress small corrective signals, and that explicitly factorizing into anchor plus residual will reliably improve geometric and contact precision without new failure modes."}},"verdict_id":"588f7b61-02b8-468f-a41d-d522fbf37258"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c688d5c98d10ef2a07db5216dfa218525feb343b0abd7f927c154c5fd949de00","target":"record","created_at":"2026-07-22T01:22:41Z","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":"17a023e370d94ad2a9fd2c3abea104c4ad14f980fa2e28df751f7b4ddc586185","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-04-20T04:25:24Z","title_canon_sha256":"7ad7e8a3d3f2c48b317c9d28cc1c8737438b7f635cb5beb9d5960ca96ef77e5a"},"schema_version":"1.0","source":{"id":"2604.17787","kind":"arxiv","version":2}},"canonical_sha256":"77bad903b3654c65d2f9dfb82efd379e5668f9ce177ed59701bc5c6fc077b86b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"77bad903b3654c65d2f9dfb82efd379e5668f9ce177ed59701bc5c6fc077b86b","first_computed_at":"2026-07-22T01:22:41.373337Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-22T01:22:41.373337Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PWRRHUCZJCCBYXraQ/kw1Rfd0CxSocj+y1vOLzg7+5Nc9lYQv8nuLRxIOeIM7hZppoNFeKzuFAKUoAB2eetjAw==","signature_status":"signed_v1","signed_at":"2026-07-22T01:22:41.374312Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.17787","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c688d5c98d10ef2a07db5216dfa218525feb343b0abd7f927c154c5fd949de00","sha256:cc042f1f98e13a601ca35132b5f2dd319f87cb4af8c3b90c09cf64c9109884ca"],"state_sha256":"d27b50abf8ed74e6fc2f5ebd1eadeee5ea30e977d64b17d823078f3650a5e5ee"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1oOaiMC1d0nezhDmjNCYB5puiGpyGC6APhUjCu1Nv9XJYc5AtHwosDYaQVUh1GTmCa6yRetHTSP0unR5/JGBBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T08:24:34.234067Z","bundle_sha256":"50e737a389a35c5c01a7575e73844c22480663b3f8f1f1911e4091c70b4bdd2d"}}