{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:J422RMZBA4C2COA5WGPNKNCNDX","short_pith_number":"pith:J422RMZB","canonical_record":{"source":{"id":"2605.13316","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:28:50Z","cross_cats_sorted":[],"title_canon_sha256":"5f4c6ea855a8375d51c624a4f7fd0324fa9e443ffa64ae58ebe7620b5942e85a","abstract_canon_sha256":"4de58a23089283a8e24f975150561e80952a4ba202fb983e0157249a78cad196"},"schema_version":"1.0"},"canonical_sha256":"4f35a8b3210705a1381db19ed5344d1dc753edf34a3e307bbb2782c16a430590","source":{"kind":"arxiv","id":"2605.13316","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13316","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13316v1","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13316","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"pith_short_12","alias_value":"J422RMZBA4C2","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"J422RMZBA4C2COA5","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"J422RMZB","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:J422RMZBA4C2COA5WGPNKNCNDX","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13316","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:28:50Z","cross_cats_sorted":[],"title_canon_sha256":"5f4c6ea855a8375d51c624a4f7fd0324fa9e443ffa64ae58ebe7620b5942e85a","abstract_canon_sha256":"4de58a23089283a8e24f975150561e80952a4ba202fb983e0157249a78cad196"},"schema_version":"1.0"},"canonical_sha256":"4f35a8b3210705a1381db19ed5344d1dc753edf34a3e307bbb2782c16a430590","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:48.771520Z","signature_b64":"wxBSnZYit9rdou4oy+fMqWPpVaaOwv1vHlL/t3r+bfIBDZRTgN2uGHjXR+T680ipppbi3lyK3DGOLFdYcf6oAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4f35a8b3210705a1381db19ed5344d1dc753edf34a3e307bbb2782c16a430590","last_reissued_at":"2026-05-18T02:44:48.771105Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:48.771105Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13316","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xQ8ANDpc1dT0r287TtYjRVFz5R5di8reHpV6Mvz2uEC2Y9np7ZrdImp3xtJLnjB7HJV6xmX8P4IqDMy6K7H8DA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-05T08:51:29.982439Z"},"content_sha256":"30621b5b13f7d0334bede6ba70d8636d7ebd18e8f99a561c20d9a38bb520c174","schema_version":"1.0","event_id":"sha256:30621b5b13f7d0334bede6ba70d8636d7ebd18e8f99a561c20d9a38bb520c174"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:J422RMZBA4C2COA5WGPNKNCNDX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Test-time Sparsity for Extreme Fast Action Diffusion","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"Test-time sparsity accelerates action diffusion 5 times with 92 percent fewer FLOPs and no performance loss.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chen Tang, Jianbo Zhou, Kangye Ji, Ye Li, Yuan Meng, Zhi Wang","submitted_at":"2026-05-13T10:28:50Z","abstract_excerpt":"Action diffusion excels at high-fidelity action generation but incurs heavy computational costs owing to its iterative denoising nature. Despite current technologies showing promise in accelerating diffusion transformers by reusing the cached features, they struggle to adapt to policy dynamics arising from diverse perceptions and multi-round rollout iterations in open environments. We propose test-time sparsity to tackle this challenge, which aims to accelerate action diffusion by dynamically predicting prunable residual computations for each model forward at test time. However, two bottleneck"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"our method reduces FLOPs by 92% and accelerates action generation by 5x, achieving lossless performance with an inference frequency of 47.5 Hz.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The omnidirectional reusing strategy and lightweight pruner can constrain large pruning errors under aggressive 95% sparsity across diverse perceptions and multi-round rollouts in open environments.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Test-time sparsity with a parallel pipeline and omnidirectional feature reuse accelerates action diffusion by 5x to 47.5 Hz while cutting FLOPs 92% with no performance loss.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Test-time sparsity accelerates action diffusion 5 times with 92 percent fewer FLOPs and no performance loss.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"94a0ee8cd5ed6bea9e639fbd0c93e8ce54eec909aaf7f8f78201c67ef12bf76a"},"source":{"id":"2605.13316","kind":"arxiv","version":1},"verdict":{"id":"7bb51af4-72e1-49f4-a689-b76020277a66","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:41:50.635837Z","strongest_claim":"our method reduces FLOPs by 92% and accelerates action generation by 5x, achieving lossless performance with an inference frequency of 47.5 Hz.","one_line_summary":"Test-time sparsity with a parallel pipeline and omnidirectional feature reuse accelerates action diffusion by 5x to 47.5 Hz while cutting FLOPs 92% with no performance loss.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The omnidirectional reusing strategy and lightweight pruner can constrain large pruning errors under aggressive 95% sparsity across diverse perceptions and multi-round rollouts in open environments.","pith_extraction_headline":"Test-time sparsity accelerates action diffusion 5 times with 92 percent fewer FLOPs and no performance loss."},"references":{"count":32,"sample":[{"doi":"","year":2024,"title":"$\\pi_0$: A Vision-Language-Action Flow Model for General Robot Control","work_id":"f790abdc-a796-482f-a40d-f8ee035ecfc2","ref_index":1,"cited_arxiv_id":"2410.24164","is_internal_anchor":true},{"doi":"","year":2025,"title":"Falcon: Fast visuomotor policies via partial denoising.arXiv preprint arXiv:2503.00339, 2025","work_id":"b6a39656-3a2c-4f0c-b774-19369dd44da9","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"arXiv preprint arXiv:2406.01125 (2024) 3, 4","work_id":"a9eb5bf3-fc76-45d5-9721-87061c21a0ba","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Diffusion policy: Visuomotor policy learning via action dif- fusion.The International Journal of Robotics Research, 44 (10-11):1684–1704","work_id":"9a143ff1-5f91-4c81-9ecd-ea644cb810ed","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2020,"title":"Relay policy learning: Solving long- horizon tasks via imitation and reinforcement learning","work_id":"068d1f8e-a0e2-4493-9660-49ab2fac0289","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":32,"snapshot_sha256":"c82330968c6f47fdda2ebbda626222e2b23834b75cb439ece74c7fdc114cbe9a","internal_anchors":9},"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":"7bb51af4-72e1-49f4-a689-b76020277a66"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UHd9rv2rT2sdhWK/vQEJ1ZIloChlu8v2HEbvTjhUxXrelcTBuKm/GH8uQZKiAvAhYGpxi9XOk/ydIHel6muODw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-05T08:51:29.983039Z"},"content_sha256":"ec29d08fd60dba74f737f9fda3d333f3e29c4d37bd82f85244519054c5b4ea21","schema_version":"1.0","event_id":"sha256:ec29d08fd60dba74f737f9fda3d333f3e29c4d37bd82f85244519054c5b4ea21"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/J422RMZBA4C2COA5WGPNKNCNDX/bundle.json","state_url":"https://pith.science/pith/J422RMZBA4C2COA5WGPNKNCNDX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/J422RMZBA4C2COA5WGPNKNCNDX/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":"2026-06-05T08:51:29Z","links":{"resolver":"https://pith.science/pith/J422RMZBA4C2COA5WGPNKNCNDX","bundle":"https://pith.science/pith/J422RMZBA4C2COA5WGPNKNCNDX/bundle.json","state":"https://pith.science/pith/J422RMZBA4C2COA5WGPNKNCNDX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/J422RMZBA4C2COA5WGPNKNCNDX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:J422RMZBA4C2COA5WGPNKNCNDX","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":"4de58a23089283a8e24f975150561e80952a4ba202fb983e0157249a78cad196","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:28:50Z","title_canon_sha256":"5f4c6ea855a8375d51c624a4f7fd0324fa9e443ffa64ae58ebe7620b5942e85a"},"schema_version":"1.0","source":{"id":"2605.13316","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13316","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13316v1","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13316","created_at":"2026-05-18T02:44:48Z"},{"alias_kind":"pith_short_12","alias_value":"J422RMZBA4C2","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"J422RMZBA4C2COA5","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"J422RMZB","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:ec29d08fd60dba74f737f9fda3d333f3e29c4d37bd82f85244519054c5b4ea21","target":"graph","created_at":"2026-05-18T02:44:48Z","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 method reduces FLOPs by 92% and accelerates action generation by 5x, achieving lossless performance with an inference frequency of 47.5 Hz."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The omnidirectional reusing strategy and lightweight pruner can constrain large pruning errors under aggressive 95% sparsity across diverse perceptions and multi-round rollouts in open environments."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Test-time sparsity with a parallel pipeline and omnidirectional feature reuse accelerates action diffusion by 5x to 47.5 Hz while cutting FLOPs 92% with no performance loss."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Test-time sparsity accelerates action diffusion 5 times with 92 percent fewer FLOPs and no performance loss."}],"snapshot_sha256":"94a0ee8cd5ed6bea9e639fbd0c93e8ce54eec909aaf7f8f78201c67ef12bf76a"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Action diffusion excels at high-fidelity action generation but incurs heavy computational costs owing to its iterative denoising nature. Despite current technologies showing promise in accelerating diffusion transformers by reusing the cached features, they struggle to adapt to policy dynamics arising from diverse perceptions and multi-round rollout iterations in open environments. We propose test-time sparsity to tackle this challenge, which aims to accelerate action diffusion by dynamically predicting prunable residual computations for each model forward at test time. However, two bottleneck","authors_text":"Chen Tang, Jianbo Zhou, Kangye Ji, Ye Li, Yuan Meng, Zhi Wang","cross_cats":[],"headline":"Test-time sparsity accelerates action diffusion 5 times with 92 percent fewer FLOPs and no performance loss.","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:28:50Z","title":"Test-time Sparsity for Extreme Fast Action Diffusion"},"references":{"count":32,"internal_anchors":9,"resolved_work":32,"sample":[{"cited_arxiv_id":"2410.24164","doi":"","is_internal_anchor":true,"ref_index":1,"title":"$\\pi_0$: A Vision-Language-Action Flow Model for General Robot Control","work_id":"f790abdc-a796-482f-a40d-f8ee035ecfc2","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Falcon: Fast visuomotor policies via partial denoising.arXiv preprint arXiv:2503.00339, 2025","work_id":"b6a39656-3a2c-4f0c-b774-19369dd44da9","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"arXiv preprint arXiv:2406.01125 (2024) 3, 4","work_id":"a9eb5bf3-fc76-45d5-9721-87061c21a0ba","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Diffusion policy: Visuomotor policy learning via action dif- fusion.The International Journal of Robotics Research, 44 (10-11):1684–1704","work_id":"9a143ff1-5f91-4c81-9ecd-ea644cb810ed","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Relay policy learning: Solving long- horizon tasks via imitation and reinforcement learning","work_id":"068d1f8e-a0e2-4493-9660-49ab2fac0289","year":2020}],"snapshot_sha256":"c82330968c6f47fdda2ebbda626222e2b23834b75cb439ece74c7fdc114cbe9a"},"source":{"id":"2605.13316","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:41:50.635837Z","id":"7bb51af4-72e1-49f4-a689-b76020277a66","model_set":{"reader":"grok-4.3"},"one_line_summary":"Test-time sparsity with a parallel pipeline and omnidirectional feature reuse accelerates action diffusion by 5x to 47.5 Hz while cutting FLOPs 92% with no performance loss.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Test-time sparsity accelerates action diffusion 5 times with 92 percent fewer FLOPs and no performance loss.","strongest_claim":"our method reduces FLOPs by 92% and accelerates action generation by 5x, achieving lossless performance with an inference frequency of 47.5 Hz.","weakest_assumption":"The omnidirectional reusing strategy and lightweight pruner can constrain large pruning errors under aggressive 95% sparsity across diverse perceptions and multi-round rollouts in open environments."}},"verdict_id":"7bb51af4-72e1-49f4-a689-b76020277a66"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:30621b5b13f7d0334bede6ba70d8636d7ebd18e8f99a561c20d9a38bb520c174","target":"record","created_at":"2026-05-18T02:44:48Z","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":"4de58a23089283a8e24f975150561e80952a4ba202fb983e0157249a78cad196","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T10:28:50Z","title_canon_sha256":"5f4c6ea855a8375d51c624a4f7fd0324fa9e443ffa64ae58ebe7620b5942e85a"},"schema_version":"1.0","source":{"id":"2605.13316","kind":"arxiv","version":1}},"canonical_sha256":"4f35a8b3210705a1381db19ed5344d1dc753edf34a3e307bbb2782c16a430590","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4f35a8b3210705a1381db19ed5344d1dc753edf34a3e307bbb2782c16a430590","first_computed_at":"2026-05-18T02:44:48.771105Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:48.771105Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"wxBSnZYit9rdou4oy+fMqWPpVaaOwv1vHlL/t3r+bfIBDZRTgN2uGHjXR+T680ipppbi3lyK3DGOLFdYcf6oAA==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:48.771520Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13316","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:30621b5b13f7d0334bede6ba70d8636d7ebd18e8f99a561c20d9a38bb520c174","sha256:ec29d08fd60dba74f737f9fda3d333f3e29c4d37bd82f85244519054c5b4ea21"],"state_sha256":"377ef5a52476032de1bbf68ca500d32401feb2f93c57dcc5f1dce7d3fe0fb558"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3wbNdmX037syhSp+jiQYTAk2FVv9GXfmgKP5EN2jzSZgqjjXsEP0QfxPj0gfQ/DnaM1LYVM5TPdves1/3QMaAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-05T08:51:29.985608Z","bundle_sha256":"63bbf842d4d6b4d44e428d29a7dfbea710d89e0f4251531926618b277d9ed801"}}