{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:EC6JEZRPDN5EFDVZNK3ZA6OCQE","short_pith_number":"pith:EC6JEZRP","canonical_record":{"source":{"id":"2603.22078","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.RO","submitted_at":"2026-03-23T15:13:15Z","cross_cats_sorted":[],"title_canon_sha256":"3849a3e9ea666495d80ce1ad88215bc9bb9bfddca30d86b13aded8233eb5a570","abstract_canon_sha256":"2ec7716672dedd3a45f9d9bcd832a1d4723d87766cccfe35353a935d31d5cbc1"},"schema_version":"1.0"},"canonical_sha256":"20bc92662f1b7a428eb96ab79079c281315a6da3fec0518d578190f9d94cd9a7","source":{"kind":"arxiv","id":"2603.22078","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2603.22078","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"arxiv_version","alias_value":"2603.22078v4","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.22078","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"pith_short_12","alias_value":"EC6JEZRPDN5E","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"pith_short_16","alias_value":"EC6JEZRPDN5EFDVZ","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"pith_short_8","alias_value":"EC6JEZRP","created_at":"2026-07-24T00:23:08Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:EC6JEZRPDN5EFDVZNK3ZA6OCQE","target":"record","payload":{"canonical_record":{"source":{"id":"2603.22078","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.RO","submitted_at":"2026-03-23T15:13:15Z","cross_cats_sorted":[],"title_canon_sha256":"3849a3e9ea666495d80ce1ad88215bc9bb9bfddca30d86b13aded8233eb5a570","abstract_canon_sha256":"2ec7716672dedd3a45f9d9bcd832a1d4723d87766cccfe35353a935d31d5cbc1"},"schema_version":"1.0"},"canonical_sha256":"20bc92662f1b7a428eb96ab79079c281315a6da3fec0518d578190f9d94cd9a7","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-24T00:23:08.870840Z","signature_b64":"r+y/eQu2trFPYIbG7ln8nmgFMZ6FotErxyUOU2M0DqPrnuj8Lu0JH3+3JFZbfmP9s7ElyfQEgW/KX33XbamEAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"20bc92662f1b7a428eb96ab79079c281315a6da3fec0518d578190f9d94cd9a7","last_reissued_at":"2026-07-24T00:23:08.869871Z","signature_status":"signed_v1","first_computed_at":"2026-07-24T00:23:08.869871Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2603.22078","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-24T00:23:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"D6qsoMnQhQGbo7IAAiGOtMjDfyajVVd4KQx0wwXOWBVPKXVmDmfU0YRrW5FJt59vYToPzR5ShkrjOhobmL39Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T19:50:29.626232Z"},"content_sha256":"848e1d759d28e1a1fd17cebda7c83793ea733bc9c598707c15fc70d456aa5300","schema_version":"1.0","event_id":"sha256:848e1d759d28e1a1fd17cebda7c83793ea733bc9c598707c15fc70d456aa5300"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:EC6JEZRPDN5EFDVZNK3ZA6OCQE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Do World Action Models Generalize Better than VLAs? A Robustness Study","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"World action models reach higher success rates than most VLAs under visual and language perturbations in robot tasks.","cross_cats":[],"primary_cat":"cs.RO","authors_text":"Amir Rasouli, Behnam Rahmati, Feng Wen, Lingfeng Zhang, Rui Heng Yang, Sajjad Pakdamansavoji, Tongtong Cao, Xingyue Quan, Xinyu Wang, Yangzheng Wu, Yingxue Zhang, Yintao Ma, Zhanguang Zhang, Zhiyuan Li","submitted_at":"2026-03-23T15:13:15Z","abstract_excerpt":"Robot action planning in the real world is challenging as it requires not only understanding the current state of the environment but also predicting how it will evolve in response to actions. Vision-language-action (VLA), which repurpose large-scale vision-language models for robot action generation using action experts, have achieved notable success across a variety of robotic tasks. Nevertheless, their performance remains constrained by the scope of their training data, exhibiting limited generalization to unseen scenarios and vulnerability to diverse contextual perturbations. More recently"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our results show that WAMs achieve strong robustness, with LingBot-VA reaching 74.2% success rate on RoboTwin 2.0-Plus and Cosmos-Policy achieving 82.2% on LIBERO-Plus.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The chosen visual and language perturbations on LIBERO-Plus and RoboTwin 2.0-Plus fairly represent the generalization challenges that matter for real-world robot deployment.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"World action models demonstrate stronger robustness to perturbations than vision-language-action models on LIBERO-Plus and RoboTwin 2.0-Plus benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"World action models reach higher success rates than most VLAs under visual and language perturbations in robot tasks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f1f7e2104f256ca58b7d34ff6ba7bfd9f78458b488215f9f52234dc6b1443b53"},"source":{"id":"2603.22078","kind":"arxiv","version":4},"verdict":{"id":"777cf6e2-7a71-4b31-b227-d7fd2c0aa0c4","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T00:36:25.060531Z","strongest_claim":"Our results show that WAMs achieve strong robustness, with LingBot-VA reaching 74.2% success rate on RoboTwin 2.0-Plus and Cosmos-Policy achieving 82.2% on LIBERO-Plus.","one_line_summary":"World action models demonstrate stronger robustness to perturbations than vision-language-action models on LIBERO-Plus and RoboTwin 2.0-Plus benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The chosen visual and language perturbations on LIBERO-Plus and RoboTwin 2.0-Plus fairly represent the generalization challenges that matter for real-world robot deployment.","pith_extraction_headline":"World action models reach higher success rates than most VLAs under visual and language perturbations in robot tasks."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2603.22078/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":2,"snapshot_sha256":"475be69147e06479717b45f92d00d2e9282c83f25b8d3fcd8cba7cb8d6990f85"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"777cf6e2-7a71-4b31-b227-d7fd2c0aa0c4"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-24T00:23:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wgvO5LTiAowBig9TA8A/GepOXZLOnar3Yr8FdGqhLQEsxSq7P4kGrvuWDMa0eUlks6M4roMJVc7qKScrBvZWAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T19:50:29.626905Z"},"content_sha256":"81ab409fefc5391d2034765a3f73e8f5ef718c14056b1b034da5ef996567a9a3","schema_version":"1.0","event_id":"sha256:81ab409fefc5391d2034765a3f73e8f5ef718c14056b1b034da5ef996567a9a3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/EC6JEZRPDN5EFDVZNK3ZA6OCQE/bundle.json","state_url":"https://pith.science/pith/EC6JEZRPDN5EFDVZNK3ZA6OCQE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/EC6JEZRPDN5EFDVZNK3ZA6OCQE/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":"2026-08-03T19:50:29Z","links":{"resolver":"https://pith.science/pith/EC6JEZRPDN5EFDVZNK3ZA6OCQE","bundle":"https://pith.science/pith/EC6JEZRPDN5EFDVZNK3ZA6OCQE/bundle.json","state":"https://pith.science/pith/EC6JEZRPDN5EFDVZNK3ZA6OCQE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/EC6JEZRPDN5EFDVZNK3ZA6OCQE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:EC6JEZRPDN5EFDVZNK3ZA6OCQE","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":"2ec7716672dedd3a45f9d9bcd832a1d4723d87766cccfe35353a935d31d5cbc1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.RO","submitted_at":"2026-03-23T15:13:15Z","title_canon_sha256":"3849a3e9ea666495d80ce1ad88215bc9bb9bfddca30d86b13aded8233eb5a570"},"schema_version":"1.0","source":{"id":"2603.22078","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2603.22078","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"arxiv_version","alias_value":"2603.22078v4","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.22078","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"pith_short_12","alias_value":"EC6JEZRPDN5E","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"pith_short_16","alias_value":"EC6JEZRPDN5EFDVZ","created_at":"2026-07-24T00:23:08Z"},{"alias_kind":"pith_short_8","alias_value":"EC6JEZRP","created_at":"2026-07-24T00:23:08Z"}],"graph_snapshots":[{"event_id":"sha256:81ab409fefc5391d2034765a3f73e8f5ef718c14056b1b034da5ef996567a9a3","target":"graph","created_at":"2026-07-24T00:23:08Z","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":"Our results show that WAMs achieve strong robustness, with LingBot-VA reaching 74.2% success rate on RoboTwin 2.0-Plus and Cosmos-Policy achieving 82.2% on LIBERO-Plus."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The chosen visual and language perturbations on LIBERO-Plus and RoboTwin 2.0-Plus fairly represent the generalization challenges that matter for real-world robot deployment."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"World action models demonstrate stronger robustness to perturbations than vision-language-action models on LIBERO-Plus and RoboTwin 2.0-Plus benchmarks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"World action models reach higher success rates than most VLAs under visual and language perturbations in robot tasks."}],"snapshot_sha256":"f1f7e2104f256ca58b7d34ff6ba7bfd9f78458b488215f9f52234dc6b1443b53"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"475be69147e06479717b45f92d00d2e9282c83f25b8d3fcd8cba7cb8d6990f85"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2603.22078/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Robot action planning in the real world is challenging as it requires not only understanding the current state of the environment but also predicting how it will evolve in response to actions. Vision-language-action (VLA), which repurpose large-scale vision-language models for robot action generation using action experts, have achieved notable success across a variety of robotic tasks. Nevertheless, their performance remains constrained by the scope of their training data, exhibiting limited generalization to unseen scenarios and vulnerability to diverse contextual perturbations. More recently","authors_text":"Amir Rasouli, Behnam Rahmati, Feng Wen, Lingfeng Zhang, Rui Heng Yang, Sajjad Pakdamansavoji, Tongtong Cao, Xingyue Quan, Xinyu Wang, Yangzheng Wu, Yingxue Zhang, Yintao Ma, Zhanguang Zhang, Zhiyuan Li","cross_cats":[],"headline":"World action models reach higher success rates than most VLAs under visual and language perturbations in robot tasks.","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.RO","submitted_at":"2026-03-23T15:13:15Z","title":"Do World Action Models Generalize Better than VLAs? A Robustness Study"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.22078","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-15T00:36:25.060531Z","id":"777cf6e2-7a71-4b31-b227-d7fd2c0aa0c4","model_set":{"reader":"grok-4.3"},"one_line_summary":"World action models demonstrate stronger robustness to perturbations than vision-language-action models on LIBERO-Plus and RoboTwin 2.0-Plus benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"World action models reach higher success rates than most VLAs under visual and language perturbations in robot tasks.","strongest_claim":"Our results show that WAMs achieve strong robustness, with LingBot-VA reaching 74.2% success rate on RoboTwin 2.0-Plus and Cosmos-Policy achieving 82.2% on LIBERO-Plus.","weakest_assumption":"The chosen visual and language perturbations on LIBERO-Plus and RoboTwin 2.0-Plus fairly represent the generalization challenges that matter for real-world robot deployment."}},"verdict_id":"777cf6e2-7a71-4b31-b227-d7fd2c0aa0c4"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:848e1d759d28e1a1fd17cebda7c83793ea733bc9c598707c15fc70d456aa5300","target":"record","created_at":"2026-07-24T00:23:08Z","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":"2ec7716672dedd3a45f9d9bcd832a1d4723d87766cccfe35353a935d31d5cbc1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.RO","submitted_at":"2026-03-23T15:13:15Z","title_canon_sha256":"3849a3e9ea666495d80ce1ad88215bc9bb9bfddca30d86b13aded8233eb5a570"},"schema_version":"1.0","source":{"id":"2603.22078","kind":"arxiv","version":4}},"canonical_sha256":"20bc92662f1b7a428eb96ab79079c281315a6da3fec0518d578190f9d94cd9a7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"20bc92662f1b7a428eb96ab79079c281315a6da3fec0518d578190f9d94cd9a7","first_computed_at":"2026-07-24T00:23:08.869871Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-24T00:23:08.869871Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"r+y/eQu2trFPYIbG7ln8nmgFMZ6FotErxyUOU2M0DqPrnuj8Lu0JH3+3JFZbfmP9s7ElyfQEgW/KX33XbamEAA==","signature_status":"signed_v1","signed_at":"2026-07-24T00:23:08.870840Z","signed_message":"canonical_sha256_bytes"},"source_id":"2603.22078","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:848e1d759d28e1a1fd17cebda7c83793ea733bc9c598707c15fc70d456aa5300","sha256:81ab409fefc5391d2034765a3f73e8f5ef718c14056b1b034da5ef996567a9a3"],"state_sha256":"868f5003da52189ad834611d87e28a035ae38b54d1f551463163e2357806fa3d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GzFl5BJsPpylyvSKYwHPe0Us+1CMure1sF3sjVKnvM173t7cB7xeOxZdCqmrlkIlkSQE5lA6pP4IJMfSF5wbDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T19:50:29.633039Z","bundle_sha256":"a7f678c53009ac6b38685ccc092d67e426b59a3d41c37dec33b5ec05dd5cd935"}}