{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:GLWKJHBICPATGCIUJRCH4FJJED","short_pith_number":"pith:GLWKJHBI","canonical_record":{"source":{"id":"2605.31598","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-29T17:59:02Z","cross_cats_sorted":[],"title_canon_sha256":"86bf6c7613a2ccd00d751c64378cb2b24eb7fbee2b8a3f2dc40d0593526f0b41","abstract_canon_sha256":"89408e460cd559186be34003c643802a2e7332c72349fc47b34c941b17cfa746"},"schema_version":"1.0"},"canonical_sha256":"32eca49c2813c13309144c447e152920ec3892166b5056e90d8a3d3001b47c48","source":{"kind":"arxiv","id":"2605.31598","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.31598","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"arxiv_version","alias_value":"2605.31598v1","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.31598","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"pith_short_12","alias_value":"GLWKJHBICPAT","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"pith_short_16","alias_value":"GLWKJHBICPATGCIU","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"pith_short_8","alias_value":"GLWKJHBI","created_at":"2026-06-01T02:04:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:GLWKJHBICPATGCIUJRCH4FJJED","target":"record","payload":{"canonical_record":{"source":{"id":"2605.31598","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-29T17:59:02Z","cross_cats_sorted":[],"title_canon_sha256":"86bf6c7613a2ccd00d751c64378cb2b24eb7fbee2b8a3f2dc40d0593526f0b41","abstract_canon_sha256":"89408e460cd559186be34003c643802a2e7332c72349fc47b34c941b17cfa746"},"schema_version":"1.0"},"canonical_sha256":"32eca49c2813c13309144c447e152920ec3892166b5056e90d8a3d3001b47c48","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T02:04:15.132561Z","signature_b64":"SUBR2IuwMWUbMKxNoquwtlsdd/D3Od+K/Hkm+Y+SgUrX3bPN2svh7LDWahyTNr8w2eUJQGZufp607qK6bvcKBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"32eca49c2813c13309144c447e152920ec3892166b5056e90d8a3d3001b47c48","last_reissued_at":"2026-06-01T02:04:15.131885Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T02:04:15.131885Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.31598","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":"2026-06-01T02:04:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"S+Lpnmli7fWpL2qIaveKXelRPG3Wn1sKglZry8TBiOc07ABX+BB1u9kV+g+8y8K/XRmUGOhJV3nEZpKSZKdeAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-06T19:00:37.247264Z"},"content_sha256":"638cd368d4d44394c85ec253e8a1dd61411cba1a582cab9ff889a18c2bed750b","schema_version":"1.0","event_id":"sha256:638cd368d4d44394c85ec253e8a1dd61411cba1a582cab9ff889a18c2bed750b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:GLWKJHBICPATGCIUJRCH4FJJED","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Linear Scaling Video VLMs for Long Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cristobal Eyzaguirre, Jiajun Wu, Juan Carlos Niebles","submitted_at":"2026-05-29T17:59:02Z","abstract_excerpt":"Video vision-language models (VLMs) are increasingly used in long-horizon and streaming settings, yet most video encoders still rely on spatiotemporal self-attention, causing compute and latency to grow quadratically with the number of frames. Existing efficiency methods improve scalability but often lose accuracy relative to full self-attention, for example through aggressive frame/token dropping or coarse attention approximations. We introduce StateKV, an inference-time method that adapts pretrained long-video VLMs to linear-time video prefill by carrying cross-frame context in a fixed-capac"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.31598","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.31598/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-01T02:04:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tqu0ptY4XPrYwB66dNv6xamfDGqEK4ipO/WI+G8V+qn4AeJzxdlQLWhhypJJ/+ZyUbVaxVZaPoPl6c3ScprjAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-06T19:00:37.248061Z"},"content_sha256":"0401cf32ff01dd1439e6c9ae16d2f5b66987f0ef78b05f538ae32b25461e2c10","schema_version":"1.0","event_id":"sha256:0401cf32ff01dd1439e6c9ae16d2f5b66987f0ef78b05f538ae32b25461e2c10"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GLWKJHBICPATGCIUJRCH4FJJED/bundle.json","state_url":"https://pith.science/pith/GLWKJHBICPATGCIUJRCH4FJJED/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GLWKJHBICPATGCIUJRCH4FJJED/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":"2026-06-06T19:00:37Z","links":{"resolver":"https://pith.science/pith/GLWKJHBICPATGCIUJRCH4FJJED","bundle":"https://pith.science/pith/GLWKJHBICPATGCIUJRCH4FJJED/bundle.json","state":"https://pith.science/pith/GLWKJHBICPATGCIUJRCH4FJJED/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GLWKJHBICPATGCIUJRCH4FJJED/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GLWKJHBICPATGCIUJRCH4FJJED","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":"89408e460cd559186be34003c643802a2e7332c72349fc47b34c941b17cfa746","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-29T17:59:02Z","title_canon_sha256":"86bf6c7613a2ccd00d751c64378cb2b24eb7fbee2b8a3f2dc40d0593526f0b41"},"schema_version":"1.0","source":{"id":"2605.31598","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.31598","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"arxiv_version","alias_value":"2605.31598v1","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.31598","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"pith_short_12","alias_value":"GLWKJHBICPAT","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"pith_short_16","alias_value":"GLWKJHBICPATGCIU","created_at":"2026-06-01T02:04:15Z"},{"alias_kind":"pith_short_8","alias_value":"GLWKJHBI","created_at":"2026-06-01T02:04:15Z"}],"graph_snapshots":[{"event_id":"sha256:0401cf32ff01dd1439e6c9ae16d2f5b66987f0ef78b05f538ae32b25461e2c10","target":"graph","created_at":"2026-06-01T02:04:15Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2605.31598/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Video vision-language models (VLMs) are increasingly used in long-horizon and streaming settings, yet most video encoders still rely on spatiotemporal self-attention, causing compute and latency to grow quadratically with the number of frames. Existing efficiency methods improve scalability but often lose accuracy relative to full self-attention, for example through aggressive frame/token dropping or coarse attention approximations. We introduce StateKV, an inference-time method that adapts pretrained long-video VLMs to linear-time video prefill by carrying cross-frame context in a fixed-capac","authors_text":"Cristobal Eyzaguirre, Jiajun Wu, Juan Carlos Niebles","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-29T17:59:02Z","title":"Linear Scaling Video VLMs for Long Video Understanding"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.31598","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:638cd368d4d44394c85ec253e8a1dd61411cba1a582cab9ff889a18c2bed750b","target":"record","created_at":"2026-06-01T02:04:15Z","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":"89408e460cd559186be34003c643802a2e7332c72349fc47b34c941b17cfa746","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-29T17:59:02Z","title_canon_sha256":"86bf6c7613a2ccd00d751c64378cb2b24eb7fbee2b8a3f2dc40d0593526f0b41"},"schema_version":"1.0","source":{"id":"2605.31598","kind":"arxiv","version":1}},"canonical_sha256":"32eca49c2813c13309144c447e152920ec3892166b5056e90d8a3d3001b47c48","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"32eca49c2813c13309144c447e152920ec3892166b5056e90d8a3d3001b47c48","first_computed_at":"2026-06-01T02:04:15.131885Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-01T02:04:15.131885Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SUBR2IuwMWUbMKxNoquwtlsdd/D3Od+K/Hkm+Y+SgUrX3bPN2svh7LDWahyTNr8w2eUJQGZufp607qK6bvcKBg==","signature_status":"signed_v1","signed_at":"2026-06-01T02:04:15.132561Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.31598","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:638cd368d4d44394c85ec253e8a1dd61411cba1a582cab9ff889a18c2bed750b","sha256:0401cf32ff01dd1439e6c9ae16d2f5b66987f0ef78b05f538ae32b25461e2c10"],"state_sha256":"271f803969af4726ffc23e3cab66ac9d9ffd6f83eab6b2406a305084f77ccd38"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tcpuqaYHP7ImyAkuv1ReKiRIkjrDwoW9s29NYrZtG052SALg9X22O/JruI9k50RH4vbtWqKwG5tYCXzHSILYDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-06T19:00:37.252128Z","bundle_sha256":"7c48ba45afe2a48f9396118b4bbfa9a593781a5ab4c27229d9213a3fe4645ae1"}}