{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:3KECOAG27EKT7EVDMBOTHH3LU3","short_pith_number":"pith:3KECOAG2","canonical_record":{"source":{"id":"2605.11817","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-12T09:08:42Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"d890d3fb46f5b83e15da28afa517f3d7a715d5bb097cd3daccd743f67ac7048e","abstract_canon_sha256":"10064b6bada4125ae37b0039636c0aaf75c085318aebb5df5037928af974560d"},"schema_version":"1.0"},"canonical_sha256":"da882700daf9153f92a3605d339f6ba6ddbabefb3cda0ee048296a78b3bbf927","source":{"kind":"arxiv","id":"2605.11817","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.11817","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.11817v2","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.11817","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"pith_short_12","alias_value":"3KECOAG27EKT","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"pith_short_16","alias_value":"3KECOAG27EKT7EVD","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"pith_short_8","alias_value":"3KECOAG2","created_at":"2026-05-20T00:04:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:3KECOAG27EKT7EVDMBOTHH3LU3","target":"record","payload":{"canonical_record":{"source":{"id":"2605.11817","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-12T09:08:42Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"d890d3fb46f5b83e15da28afa517f3d7a715d5bb097cd3daccd743f67ac7048e","abstract_canon_sha256":"10064b6bada4125ae37b0039636c0aaf75c085318aebb5df5037928af974560d"},"schema_version":"1.0"},"canonical_sha256":"da882700daf9153f92a3605d339f6ba6ddbabefb3cda0ee048296a78b3bbf927","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:04:36.179164Z","signature_b64":"qspYdYi+6kwO7am0kfFS9StvP19hj4j+aXBEgZ6zOlGHnFZ0kCC7o/Q2Y5ivQdqzZpuWggayqgbQp4d9Q/gkDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"da882700daf9153f92a3605d339f6ba6ddbabefb3cda0ee048296a78b3bbf927","last_reissued_at":"2026-05-20T00:04:36.178020Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:04:36.178020Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.11817","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":"2026-05-20T00:04:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"g41QA7kLFxe21lcUKwLIqd55/QL6dVAvoIvXWqevQi5z1tnbcudaYCfeWb7NsoGjLp8F2mCEjg0eY71W0hOrCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T22:33:51.088484Z"},"content_sha256":"a4b2d548dac0ee4f3b5523a7ae8e394663f372b0a9903edeb00bbc7681007783","schema_version":"1.0","event_id":"sha256:a4b2d548dac0ee4f3b5523a7ae8e394663f372b0a9903edeb00bbc7681007783"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:3KECOAG27EKT7EVDMBOTHH3LU3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Differentiable grid sampling lets VLA models compress visual tokens to under 10 percent while keeping full manipulation success.","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Chang Xu, Chengbin Du, Chenghao Xia, Yanxiang Ma, Yixu Feng, Yunke Wang, Zinan Zhao","submitted_at":"2026-05-12T09:08:42Z","abstract_excerpt":"Vision-Language-Action (VLA) models have shown remarkable promise in robotics manipulation, yet their high computational cost hinders real-time deployment. Existing token pruning methods suffer from a fundamental trade-off: aggressive compression using pruning inevitably discards critical geometric details like contact points, leading to severe performance degradation. This forces a compromise, limiting the achievable compression rate and thus the potential speedup. We argue that breaking this trade-off requires rethinking compression as a geometry-aware, continuous token resampling in the vis"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"validating the lowest feasible visual token count reported to date, GridS achieves a 76% reduction in FLOPs with no degradation in the success rate.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That adaptively predicting a minimal set of salient coordinates and extracting features via differentiable interpolation will always preserve the critical geometric details (contact points, spatial relations) required for manipulation success across tasks and environments.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"GridS reduces visual tokens in VLA models to under 10% of the original count via task-aware differentiable resampling, delivering 76% lower FLOPs with no drop in task success rate on benchmarks and real robots.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Differentiable grid sampling lets VLA models compress visual tokens to under 10 percent while keeping full manipulation success.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"b7a26ca9e662d38a871a961c34d9c9af25335b81066568a1e799cdd4343096e4"},"source":{"id":"2605.11817","kind":"arxiv","version":2},"verdict":{"id":"02473b66-5ced-4ed6-aa02-292631ff9758","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-13T05:37:23.048495Z","strongest_claim":"validating the lowest feasible visual token count reported to date, GridS achieves a 76% reduction in FLOPs with no degradation in the success rate.","one_line_summary":"GridS reduces visual tokens in VLA models to under 10% of the original count via task-aware differentiable resampling, delivering 76% lower FLOPs with no drop in task success rate on benchmarks and real robots.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That adaptively predicting a minimal set of salient coordinates and extracting features via differentiable interpolation will always preserve the critical geometric details (contact points, spatial relations) required for manipulation success across tasks and environments.","pith_extraction_headline":"Differentiable grid sampling lets VLA models compress visual tokens to under 10 percent while keeping full manipulation success."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.11817/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T11:37:09.344165Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T09:01:18.728167Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T08:09:33.258817Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"5369e7ac0214a6872f8302180b02f9bfb6ef7fedf5ac01a1556e47fe9a8a1d7f"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":1,"snapshot_sha256":"bc3aa9c3266c2b61e4fe9b2fd3c9ad46fd5fe3dd562c25bf4020765d7a7ffc98"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"02473b66-5ced-4ed6-aa02-292631ff9758"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:04:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"M8yRrqFDEFPUcy8WMEgVMSWgiA/NXewSe4P0+m7IvCRTfY/EflK/8WlgQj7/JXF2/sjz5eWDUK1qybzeTn/UAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T22:33:51.089377Z"},"content_sha256":"0bdbeb3c1955e51f76ebe54b6a0a360f5301b4b44717448c7ab3b8d8d920f3d2","schema_version":"1.0","event_id":"sha256:0bdbeb3c1955e51f76ebe54b6a0a360f5301b4b44717448c7ab3b8d8d920f3d2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/3KECOAG27EKT7EVDMBOTHH3LU3/bundle.json","state_url":"https://pith.science/pith/3KECOAG27EKT7EVDMBOTHH3LU3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/3KECOAG27EKT7EVDMBOTHH3LU3/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":"2026-05-26T22:33:51Z","links":{"resolver":"https://pith.science/pith/3KECOAG27EKT7EVDMBOTHH3LU3","bundle":"https://pith.science/pith/3KECOAG27EKT7EVDMBOTHH3LU3/bundle.json","state":"https://pith.science/pith/3KECOAG27EKT7EVDMBOTHH3LU3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/3KECOAG27EKT7EVDMBOTHH3LU3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:3KECOAG27EKT7EVDMBOTHH3LU3","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":"10064b6bada4125ae37b0039636c0aaf75c085318aebb5df5037928af974560d","cross_cats_sorted":["cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-12T09:08:42Z","title_canon_sha256":"d890d3fb46f5b83e15da28afa517f3d7a715d5bb097cd3daccd743f67ac7048e"},"schema_version":"1.0","source":{"id":"2605.11817","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.11817","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.11817v2","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.11817","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"pith_short_12","alias_value":"3KECOAG27EKT","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"pith_short_16","alias_value":"3KECOAG27EKT7EVD","created_at":"2026-05-20T00:04:36Z"},{"alias_kind":"pith_short_8","alias_value":"3KECOAG2","created_at":"2026-05-20T00:04:36Z"}],"graph_snapshots":[{"event_id":"sha256:0bdbeb3c1955e51f76ebe54b6a0a360f5301b4b44717448c7ab3b8d8d920f3d2","target":"graph","created_at":"2026-05-20T00:04:36Z","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":"validating the lowest feasible visual token count reported to date, GridS achieves a 76% reduction in FLOPs with no degradation in the success rate."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That adaptively predicting a minimal set of salient coordinates and extracting features via differentiable interpolation will always preserve the critical geometric details (contact points, spatial relations) required for manipulation success across tasks and environments."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"GridS reduces visual tokens in VLA models to under 10% of the original count via task-aware differentiable resampling, delivering 76% lower FLOPs with no drop in task success rate on benchmarks and real robots."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Differentiable grid sampling lets VLA models compress visual tokens to under 10 percent while keeping full manipulation success."}],"snapshot_sha256":"b7a26ca9e662d38a871a961c34d9c9af25335b81066568a1e799cdd4343096e4"},"formal_canon":{"evidence_count":1,"snapshot_sha256":"bc3aa9c3266c2b61e4fe9b2fd3c9ad46fd5fe3dd562c25bf4020765d7a7ffc98"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T11:37:09.344165Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T09:01:18.728167Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T08:09:33.258817Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.11817/integrity.json","findings":[],"snapshot_sha256":"5369e7ac0214a6872f8302180b02f9bfb6ef7fedf5ac01a1556e47fe9a8a1d7f","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Vision-Language-Action (VLA) models have shown remarkable promise in robotics manipulation, yet their high computational cost hinders real-time deployment. Existing token pruning methods suffer from a fundamental trade-off: aggressive compression using pruning inevitably discards critical geometric details like contact points, leading to severe performance degradation. This forces a compromise, limiting the achievable compression rate and thus the potential speedup. We argue that breaking this trade-off requires rethinking compression as a geometry-aware, continuous token resampling in the vis","authors_text":"Chang Xu, Chengbin Du, Chenghao Xia, Yanxiang Ma, Yixu Feng, Yunke Wang, Zinan Zhao","cross_cats":["cs.CV"],"headline":"Differentiable grid sampling lets VLA models compress visual tokens to under 10 percent while keeping full manipulation success.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-12T09:08:42Z","title":"See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Model"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.11817","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-13T05:37:23.048495Z","id":"02473b66-5ced-4ed6-aa02-292631ff9758","model_set":{"reader":"grok-4.3"},"one_line_summary":"GridS reduces visual tokens in VLA models to under 10% of the original count via task-aware differentiable resampling, delivering 76% lower FLOPs with no drop in task success rate on benchmarks and real robots.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Differentiable grid sampling lets VLA models compress visual tokens to under 10 percent while keeping full manipulation success.","strongest_claim":"validating the lowest feasible visual token count reported to date, GridS achieves a 76% reduction in FLOPs with no degradation in the success rate.","weakest_assumption":"That adaptively predicting a minimal set of salient coordinates and extracting features via differentiable interpolation will always preserve the critical geometric details (contact points, spatial relations) required for manipulation success across tasks and environments."}},"verdict_id":"02473b66-5ced-4ed6-aa02-292631ff9758"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:a4b2d548dac0ee4f3b5523a7ae8e394663f372b0a9903edeb00bbc7681007783","target":"record","created_at":"2026-05-20T00:04:36Z","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":"10064b6bada4125ae37b0039636c0aaf75c085318aebb5df5037928af974560d","cross_cats_sorted":["cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-12T09:08:42Z","title_canon_sha256":"d890d3fb46f5b83e15da28afa517f3d7a715d5bb097cd3daccd743f67ac7048e"},"schema_version":"1.0","source":{"id":"2605.11817","kind":"arxiv","version":2}},"canonical_sha256":"da882700daf9153f92a3605d339f6ba6ddbabefb3cda0ee048296a78b3bbf927","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"da882700daf9153f92a3605d339f6ba6ddbabefb3cda0ee048296a78b3bbf927","first_computed_at":"2026-05-20T00:04:36.178020Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:04:36.178020Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"qspYdYi+6kwO7am0kfFS9StvP19hj4j+aXBEgZ6zOlGHnFZ0kCC7o/Q2Y5ivQdqzZpuWggayqgbQp4d9Q/gkDw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:04:36.179164Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.11817","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a4b2d548dac0ee4f3b5523a7ae8e394663f372b0a9903edeb00bbc7681007783","sha256:0bdbeb3c1955e51f76ebe54b6a0a360f5301b4b44717448c7ab3b8d8d920f3d2"],"state_sha256":"ffb89b6642514e51d36cc1683157dcb99e5848338e5156668d1cbc2c0b314feb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6hgaz58xpt+K1Iox5S91ZecVYHB1RYXcTWY6N+3ETLgqvkK9Ulf3XivKkOLDo54p5dHpbaaYKaMWfnm9I8tkCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-26T22:33:51.092854Z","bundle_sha256":"3518f7d45f9ea08b870138329db7bc2eafb4fbc5c0db9ad66f2b5b051254262f"}}