{"id":"5a55d88d2086d1759d0f5a9ae9dd565fefe2ec4b2719a435130c4a93b46ca193","title":"Xiaoyang Xing and coauthors introduce APS-VLA for faster robotic control","argument":"Xiaoyang Xing and coauthors introduce APS-VLA, a way to reduce the visual information a vision-language-action (VLA) policy processes before generating robot actions. Its Decision Influence Estimator selects tokens using current visual observations, language instructions and robot proprioception, and is trained on robot demonstrations. The pretrained VLA backbone stays frozen; at deployment, the estimator selects and packs tokens in one step, and an auxiliary predictor used during training is removed. The authors report experiments with OpenVLA-OFT on LIBERO, and with π0.5 and XR-1 on RoboCasa365. Across five real-world tasks on two dual-arm platforms, they report mean policy inference latency falling from 137.25 to 93.25 ms, a 1.47× speedup, while average success was 61%, compared with 60% for the unpruned baseline. Escape Meridian interpretation: in these evaluations, the reported speedup came with a one-percentage-point difference in aggregate success, while the policy parameters and native action-generation procedure stayed unchanged. The abstract reports results from these evaluations, not a guarantee of similar performance elsewhere. It gives aggregate success rates, not the rate for each real-world task.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The reported real-world result covers five tasks on two dual-arm platforms; the abstract does not establish performance beyond those evaluations. The abstract reports aggregate success rates, not task-by-task outcomes.","speaker":{"name":"Xing, Xiaoyang, Feng, Chun-Mei, Cohn, Anthony, Gong, Xiaoliang, Nandi, Asoke","roleAtSource":"Article author"},"source":{"title":"APS-VLA: Action-Guided and Future-Aware Token Pruning for Efficient Vision-Language-Action Models","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.13030","publishedAt":"2026-11-09","dateBasis":"Date basis not provided."},"evidence":[{"text":"APS-VLA learns to select visual tokens from robot demonstrations while leaving the pretrained VLA backbone frozen.","excerpt":"We introduce APS-VLA, a framework that learns visual token selection from robot demonstrations while keeping the pretrained VLA backbone frozen.","url":"https://arxiv.org/abs/2611.13030","relationship":"Source-stated"},{"text":"The Decision Influence Estimator scores tokens using current visual observations, language instructions and robot proprioception.","excerpt":"A lightweight Decision Influence Estimator scores tokens conditioned on current visual observations, language instructions, and robot proprioception.","url":"https://arxiv.org/abs/2611.13030","relationship":"Source-stated"},{"text":"The authors report a reduction in mean policy inference latency, with average success close to the unpruned baseline, across five real-world tasks on two dual-arm platforms.","excerpt":"Across five real-world tasks on two dual-arm platforms, APS-VLA reduces mean policy inference latency from 137.25 to 93.25 ms, achieving a $1.47\\times$ speedup with an average success rate of 61%, compared with 60% for the unpruned baseline.","url":"https://arxiv.org/abs/2611.13030","relationship":"Source-stated"},{"text":"The abstract reports aggregate success for the five real-world tasks but does not provide the rate for each task.","excerpt":"with an average success rate of 61%, compared with 60% for the unpruned baseline.","url":"https://arxiv.org/abs/2611.13030","relationship":"Escape Meridian interpretation"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.13030","checkedAt":"2026-11-12T13:06:10.005Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What were the success rates on each of the five real-world tasks, and how did they compare with the unpruned baseline?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Xiaoyang Xing and coauthors introduce APS-VLA for faster robotic control","summary":"Xiaoyang Xing and coauthors introduce APS-VLA, a way to reduce the visual information a vision-language-action (VLA) policy processes before generating robot actions. Its Decision Influence Estimator selects tokens using current visual observations, language instructions and robot proprioception, and is trained on robot demonstrations. The pretrained VLA backbone stays frozen; at deployment, the estimator selects and packs tokens in one step, and an auxiliary predictor used during training is removed. The authors report experiments with OpenVLA-OFT on LIBERO, and with π0.5 and XR-1 on RoboCasa365. Across five real-world tasks on two dual-arm platforms, they report mean policy inference latency falling from 137.25 to 93.25 ms, a 1.47× speedup, while average success was 61%, compared with 60% for the unpruned baseline. Escape Meridian interpretation: in these evaluations, the reported speedup came with a one-percentage-point difference in aggregate success, while the policy parameters and native action-generation procedure stayed unchanged. The abstract reports results from these evaluations, not a guarantee of similar performance elsewhere. It gives aggregate success rates, not the rate for each real-world task.","question":"What were the success rates on each of the five real-world tasks, and how did they compare with the unpruned baseline?","claims":[{"text":"APS-VLA learns to select visual tokens from robot demonstrations while leaving the pretrained VLA backbone frozen.","excerpt":"We introduce APS-VLA, a framework that learns visual token selection from robot demonstrations while keeping the pretrained VLA backbone frozen.","support":"source-stated"},{"text":"The Decision Influence Estimator scores tokens using current visual observations, language instructions and robot proprioception.","excerpt":"A lightweight Decision Influence Estimator scores tokens conditioned on current visual observations, language instructions, and robot proprioception.","support":"source-stated"},{"text":"The authors report a reduction in mean policy inference latency, with average success close to the unpruned baseline, across five real-world tasks on two dual-arm platforms.","excerpt":"Across five real-world tasks on two dual-arm platforms, APS-VLA reduces mean policy inference latency from 137.25 to 93.25 ms, achieving a $1.47\\times$ speedup with an average success rate of 61%, compared with 60% for the unpruned baseline.","support":"source-stated"},{"text":"The abstract reports aggregate success for the five real-world tasks but does not provide the rate for each task.","excerpt":"with an average success rate of 61%, compared with 60% for the unpruned baseline.","support":"interpretation"}],"limitations":["The reported real-world result covers five tasks on two dual-arm platforms; the abstract does not establish performance beyond those evaluations.","The abstract reports aggregate success rates, not task-by-task outcomes."],"mediaId":"f18918152d7fe67dad86d40e066dacbf27aee9ab091f054f960eeec640d68bba","correctionReason":null,"id":"5a55d88d2086d1759d0f5a9ae9dd565fefe2ec4b2719a435130c4a93b46ca193","publicationId":"0a1655cf-a502-4bd4-8c7f-4c0812b94519","version":1,"sourceUrl":"https://arxiv.org/abs/2611.13030","sourceTitle":"APS-VLA: Action-Guided and Future-Aware Token Pruning for Efficient Vision-Language-Action Models","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-09","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-11T05:06:39.318Z","createdAt":"2026-11-11T20:02:02.528Z","historical":false,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"f18918152d7fe67dad86d40e066dacbf27aee9ab091f054f960eeec640d68bba","kind":"article","url":"https://arxiv.org/abs/2611.13030","sourceUrl":"https://arxiv.org/abs/2611.13030","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-11T20:00:57.298Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.13030","checkedAt":"2026-11-12T13:06:10.005Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"f18918152d7fe67dad86d40e066dacbf27aee9ab091f054f960eeec640d68bba","confidence":0.93,"provider":"typesafe","model":"jev-1.13.0","inputHash":"3e5e0fca0d250984e4ce182924bd35cbb2b3d6b4491290c6afd8e48ab05fc142","decidedAt":"2026-11-11T20:00:59.578Z"},"provenance":{"textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observationId":"7b234736cddd6a3ce00fd05d08efd1e7f6af9e9d3ce7bc8a5b9e440a926e9d8e","retrievedAt":"2026-11-11T20:00:56.972Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"95efa567-5ef5-4d90-8452-86746382a577"},"reviewer":{"model":"gpt-6-luna","runId":"6c0dd550-bcd5-49b1-86a0-ddf86ffb6ddb"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-11T20:02:02.528Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Xing, Xiaoyang","url":null,"kind":"unknown"},{"name":"Feng, Chun-Mei","url":null,"kind":"unknown"},{"name":"Cohn, Anthony","url":null,"kind":"unknown"},{"name":"Gong, Xiaoliang","url":null,"kind":"unknown"},{"name":"Nandi, Asoke","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Xing, Xiaoyang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Xing, Xiaoyang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}},{"actor":{"name":"Feng, Chun-Mei","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Feng, Chun-Mei","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}},{"actor":{"name":"Cohn, Anthony","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cohn, Anthony","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}},{"actor":{"name":"Gong, Xiaoliang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Gong, Xiaoliang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}},{"actor":{"name":"Nandi, Asoke","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Nandi, Asoke","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}}],"participationCoverage":{"total":5,"returned":5,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.13030","snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z","basis":"publisher-page-metadata","sourceLinks":43,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/5a55d88d2086d1759d0f5a9ae9dd565fefe2ec4b2719a435130c4a93b46ca193/source"}},"canonicalUrl":"https://escapemeridian.com/readings/xiaoyang-xing-and-coauthors-introduce-aps-vla-for-faster-5a55d88d","formats":{"html":"/readings/xiaoyang-xing-and-coauthors-introduce-aps-vla-for-faster-5a55d88d","markdown":"/readings/xiaoyang-xing-and-coauthors-introduce-aps-vla-for-faster-5a55d88d.md","json":"/readings/xiaoyang-xing-and-coauthors-introduce-aps-vla-for-faster-5a55d88d.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-pre-training","name":"Pre-training","explanation":"A model's initial training on a large dataset, before further adaptation for particular uses.","referenceUrl":"https://developers.google.com/machine-learning/glossary#pre-training"},{"id":"term-latency","name":"Latency","explanation":"The delay between sending a request and receiving the result.","referenceUrl":"https://developers.google.com/machine-learning/glossary#latency"}],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"5a55d88d2086d1759d0f5a9ae9dd565fefe2ec4b2719a435130c4a93b46ca193","revision":"0a1655cf-a502-4bd4-8c7f-4c0812b94519","revisionKind":"publication-id","title":"Xiaoyang Xing and coauthors introduce APS-VLA for faster robotic control","url":"https://escapemeridian.com/readings/xiaoyang-xing-and-coauthors-introduce-aps-vla-for-faster-5a55d88d"},"source":{"type":"source","id":"https://arxiv.org/abs/2611.13030","revision":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","revisionKind":"sha256-source-text","title":"APS-VLA: Action-Guided and Future-Aware Token Pruning for Efficient Vision-Language-Action Models","url":"https://arxiv.org/abs/2611.13030"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"5a55d88d2086d1759d0f5a9ae9dd565fefe2ec4b2719a435130c4a93b46ca193","title":"Xiaoyang Xing and coauthors introduce APS-VLA for faster robotic control","argument":"Xiaoyang Xing and coauthors introduce APS-VLA, a way to reduce the visual information a vision-language-action (VLA) policy processes before generating robot actions. Its Decision Influence Estimator selects tokens using current visual observations, language instructions and robot proprioception, and is trained on robot demonstrations. The pretrained VLA backbone stays frozen; at deployment, the estimator selects and packs tokens in one step, and an auxiliary predictor used during training is removed. The authors report experiments with OpenVLA-OFT on LIBERO, and with π0.5 and XR-1 on RoboCasa365. Across five real-world tasks on two dual-arm platforms, they report mean policy inference latency falling from 137.25 to 93.25 ms, a 1.47× speedup, while average success was 61%, compared with 60% for the unpruned baseline. Escape Meridian interpretation: in these evaluations, the reported speedup came with a one-percentage-point difference in aggregate success, while the policy parameters and native action-generation procedure stayed unchanged. The abstract reports results from these evaluations, not a guarantee of similar performance elsewhere. It gives aggregate success rates, not the rate for each real-world task.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The reported real-world result covers five tasks on two dual-arm platforms; the abstract does not establish performance beyond those evaluations. The abstract reports aggregate success rates, not task-by-task outcomes.","speaker":{"name":"Xing, Xiaoyang, Feng, Chun-Mei, Cohn, Anthony, Gong, Xiaoliang, Nandi, Asoke","roleAtSource":"Article author"},"source":{"title":"APS-VLA: Action-Guided and Future-Aware Token Pruning for Efficient Vision-Language-Action Models","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.13030","publishedAt":"2026-11-09","dateBasis":"Date basis not provided."},"evidence":[{"text":"APS-VLA learns to select visual tokens from robot demonstrations while leaving the pretrained VLA backbone frozen.","excerpt":"We introduce APS-VLA, a framework that learns visual token selection from robot demonstrations while keeping the pretrained VLA backbone frozen.","url":"https://arxiv.org/abs/2611.13030","relationship":"Source-stated"},{"text":"The Decision Influence Estimator scores tokens using current visual observations, language instructions and robot proprioception.","excerpt":"A lightweight Decision Influence Estimator scores tokens conditioned on current visual observations, language instructions, and robot proprioception.","url":"https://arxiv.org/abs/2611.13030","relationship":"Source-stated"},{"text":"The authors report a reduction in mean policy inference latency, with average success close to the unpruned baseline, across five real-world tasks on two dual-arm platforms.","excerpt":"Across five real-world tasks on two dual-arm platforms, APS-VLA reduces mean policy inference latency from 137.25 to 93.25 ms, achieving a $1.47\\times$ speedup with an average success rate of 61%, compared with 60% for the unpruned baseline.","url":"https://arxiv.org/abs/2611.13030","relationship":"Source-stated"},{"text":"The abstract reports aggregate success for the five real-world tasks but does not provide the rate for each task.","excerpt":"with an average success rate of 61%, compared with 60% for the unpruned baseline.","url":"https://arxiv.org/abs/2611.13030","relationship":"Escape Meridian interpretation"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.13030","checkedAt":"2026-11-12T13:06:10.005Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What were the success rates on each of the five real-world tasks, and how did they compare with the unpruned baseline?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Xiaoyang Xing and coauthors introduce APS-VLA for faster robotic control","summary":"Xiaoyang Xing and coauthors introduce APS-VLA, a way to reduce the visual information a vision-language-action (VLA) policy processes before generating robot actions. Its Decision Influence Estimator selects tokens using current visual observations, language instructions and robot proprioception, and is trained on robot demonstrations. The pretrained VLA backbone stays frozen; at deployment, the estimator selects and packs tokens in one step, and an auxiliary predictor used during training is removed. The authors report experiments with OpenVLA-OFT on LIBERO, and with π0.5 and XR-1 on RoboCasa365. Across five real-world tasks on two dual-arm platforms, they report mean policy inference latency falling from 137.25 to 93.25 ms, a 1.47× speedup, while average success was 61%, compared with 60% for the unpruned baseline. Escape Meridian interpretation: in these evaluations, the reported speedup came with a one-percentage-point difference in aggregate success, while the policy parameters and native action-generation procedure stayed unchanged. The abstract reports results from these evaluations, not a guarantee of similar performance elsewhere. It gives aggregate success rates, not the rate for each real-world task.","question":"What were the success rates on each of the five real-world tasks, and how did they compare with the unpruned baseline?","claims":[{"text":"APS-VLA learns to select visual tokens from robot demonstrations while leaving the pretrained VLA backbone frozen.","excerpt":"We introduce APS-VLA, a framework that learns visual token selection from robot demonstrations while keeping the pretrained VLA backbone frozen.","support":"source-stated"},{"text":"The Decision Influence Estimator scores tokens using current visual observations, language instructions and robot proprioception.","excerpt":"A lightweight Decision Influence Estimator scores tokens conditioned on current visual observations, language instructions, and robot proprioception.","support":"source-stated"},{"text":"The authors report a reduction in mean policy inference latency, with average success close to the unpruned baseline, across five real-world tasks on two dual-arm platforms.","excerpt":"Across five real-world tasks on two dual-arm platforms, APS-VLA reduces mean policy inference latency from 137.25 to 93.25 ms, achieving a $1.47\\times$ speedup with an average success rate of 61%, compared with 60% for the unpruned baseline.","support":"source-stated"},{"text":"The abstract reports aggregate success for the five real-world tasks but does not provide the rate for each task.","excerpt":"with an average success rate of 61%, compared with 60% for the unpruned baseline.","support":"interpretation"}],"limitations":["The reported real-world result covers five tasks on two dual-arm platforms; the abstract does not establish performance beyond those evaluations.","The abstract reports aggregate success rates, not task-by-task outcomes."],"mediaId":"f18918152d7fe67dad86d40e066dacbf27aee9ab091f054f960eeec640d68bba","correctionReason":null,"id":"5a55d88d2086d1759d0f5a9ae9dd565fefe2ec4b2719a435130c4a93b46ca193","publicationId":"0a1655cf-a502-4bd4-8c7f-4c0812b94519","version":1,"sourceUrl":"https://arxiv.org/abs/2611.13030","sourceTitle":"APS-VLA: Action-Guided and Future-Aware Token Pruning for Efficient Vision-Language-Action Models","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-09","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-11T05:06:39.318Z","createdAt":"2026-11-11T20:02:02.528Z","historical":false,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"f18918152d7fe67dad86d40e066dacbf27aee9ab091f054f960eeec640d68bba","kind":"article","url":"https://arxiv.org/abs/2611.13030","sourceUrl":"https://arxiv.org/abs/2611.13030","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-11T20:00:57.298Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.13030","checkedAt":"2026-11-12T13:06:10.005Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"f18918152d7fe67dad86d40e066dacbf27aee9ab091f054f960eeec640d68bba","confidence":0.93,"provider":"typesafe","model":"jev-1.13.0","inputHash":"3e5e0fca0d250984e4ce182924bd35cbb2b3d6b4491290c6afd8e48ab05fc142","decidedAt":"2026-11-11T20:00:59.578Z"},"provenance":{"textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observationId":"7b234736cddd6a3ce00fd05d08efd1e7f6af9e9d3ce7bc8a5b9e440a926e9d8e","retrievedAt":"2026-11-11T20:00:56.972Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"95efa567-5ef5-4d90-8452-86746382a577"},"reviewer":{"model":"gpt-6-luna","runId":"6c0dd550-bcd5-49b1-86a0-ddf86ffb6ddb"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-11T20:02:02.528Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Xing, Xiaoyang","url":null,"kind":"unknown"},{"name":"Feng, Chun-Mei","url":null,"kind":"unknown"},{"name":"Cohn, Anthony","url":null,"kind":"unknown"},{"name":"Gong, Xiaoliang","url":null,"kind":"unknown"},{"name":"Nandi, Asoke","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Xing, Xiaoyang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Xing, Xiaoyang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}},{"actor":{"name":"Feng, Chun-Mei","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Feng, Chun-Mei","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}},{"actor":{"name":"Cohn, Anthony","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cohn, Anthony","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}},{"actor":{"name":"Gong, Xiaoliang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Gong, Xiaoliang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}},{"actor":{"name":"Nandi, Asoke","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Nandi, Asoke","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13030","locator":"head/meta[name=author|citation_author]"},"assertionId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T13:06:10.005Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","sourceUrl":"https://arxiv.org/abs/2611.13030","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z"}}],"participationCoverage":{"total":5,"returned":5,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.13030","snapshotId":"4e2402139c9cbe745e8de5e258880d8ff1cf70f6fe1f817aeb943441b91b61ae","contentSha256":"5145cd28f21a152706393dff3d4e8543f2e104bd691492f80fb7f91ff3b2bd89","textHash":"e3525187170d2f7cfd482f06550f3199213ad0fed207add72ed3fa501ec575a5","observedAt":"2026-11-12T13:06:10.005Z","basis":"publisher-page-metadata","sourceLinks":43,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/5a55d88d2086d1759d0f5a9ae9dd565fefe2ec4b2719a435130c4a93b46ca193/source"}}}],"audits":[{"publicationId":"0a1655cf-a502-4bd4-8c7f-4c0812b94519","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading attributes APS-VLA and its reported results to Xiaoyang Xing and coauthors. This matches the arXiv author list, and the results stay framed as author reports rather than established fact.","excerpt":"Authors:Xiaoyang Xing, Chun-Mei Feng, Anthony Cohn, Xiaoliang Gong, Asoke Nandi"},{"dimension":"summary","status":"uncertain","explanation":"The summary names the benchmark experiments (OpenVLA-OFT on LIBERO; π0.5 and XR-1 on RoboCasa365) but leaves out their reported outcome: acceleration with modest changes in task success. Only the favorable real-world result (61% vs 60%) is quantified, so readers may infer success was unaffected everywhere. Smallest repair: add that the authors report inference acceleration with modest changes in task success on these benchmarks. The reading also never says plainly why lower latency matters for robot control. Terms such as token, proprioception and inference latency should get term-library tooltips.","excerpt":"demonstrate inference acceleration with modest changes in task success"},{"dimension":"date","status":"pass","explanation":"originalPublished is 2026-11-09. This matches the arXiv v1 submission date, not the later feed date of 2026-11-11.","excerpt":"[Submitted on 9 Nov 2026]"},{"dimension":"support","status":"pass","explanation":"The claims, latency figures, speedup and success rates match the abstract. The frozen backbone, the removed auxiliary predictor and the unchanged policy parameters are all stated in the source. The Escape Meridian interpretation is labeled and limited to the reported evaluations. The question about per-task success follows from the abstract giving only averages, without implying an answer.","excerpt":"APS-VLA reduces mean policy inference latency from 137.25 to 93.25 ms, achieving a $1.47\\times$ speedup with an average success rate of 61%, compared with 60% for the unpruned baseline"},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct link to the original arXiv abstract page. It matches the source URL and is the most direct way to reach the paper.","excerpt":null}],"createdAt":"2026-11-11T20:02:02.528Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}