{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:KL45WMUMJ2W6BGCCHJPPEVL3U5","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":"831cda8484bd628f8a5135b3999062c1115841ad209086668e028b944115fa15","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-14T03:24:09Z","title_canon_sha256":"28ec58e343c854ab448e2b5a731c91a40049bf1c163053032dd5760388681c72"},"schema_version":"1.0","source":{"id":"2605.14315","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.14315","created_at":"2026-05-17T23:39:09Z"},{"alias_kind":"arxiv_version","alias_value":"2605.14315v1","created_at":"2026-05-17T23:39:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14315","created_at":"2026-05-17T23:39:09Z"},{"alias_kind":"pith_short_12","alias_value":"KL45WMUMJ2W6","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"KL45WMUMJ2W6BGCC","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"KL45WMUM","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:6eaa0846f616c9d8a8297dda0876be1fc6954b22f0d79ce736d0e5441bd5252b","target":"graph","created_at":"2026-05-17T23:39:09Z","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":"TurboVGGT achieves fast multi-view reconstruction while maintaining competitive reconstruction quality compared with state-of-the-art methods."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The assumption that adaptively learning representative tokens with varying sparsity levels across frames, layers, and structurally informative regions will reliably capture global relationships without losing critical geometric details in diverse real-world scenes."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"TurboVGGT uses adaptive sparse global attention with varying sparsity levels across frames and layers plus frame attention to enable faster multi-view 3D reconstruction while keeping competitive quality versus prior state-of-the-art methods."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"TurboVGGT speeds multi-view 3D reconstruction by learning varying sparsity in attention across frames and layers."}],"snapshot_sha256":"001e251d70065c3bb979959c93fbe5ddf0638b89e0e9e9d037e3014ad20156eb"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Recent feed-forward 3D reconstruction methods, such as visual geometry transformers, have substantially advanced the traditional per-scene optimization paradigm by enabling effective multi-view reconstruction in a single forward pass. However, most existing methods struggle to achieve a balance between reconstruction quality and computational efficiency, which limits their scalability and efficiency. Although some efficient visual geometry transformers have recently emerged, they typically use the same sparsity ratio across layers and frames and lack mechanisms to adaptively learn representati","authors_text":"Bingbing Liu, Chengjie Huang, David Huang, Dongfeng Bai, Guile Wu","cross_cats":[],"headline":"TurboVGGT speeds multi-view 3D reconstruction by learning varying sparsity in attention across frames and layers.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-14T03:24:09Z","title":"TurboVGGT: Fast Visual Geometry Reconstruction with Adaptive Alternating Attention"},"references":{"count":54,"internal_anchors":2,"resolved_work":54,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Mapillary planet-scale depth dataset","work_id":"4b55c119-e053-4b1f-9c10-30689584b3f1","year":2020},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Scene- script: Reconstructing scenes with an autoregressive structured language model","work_id":"4a0199e9-ec3d-4806-b69f-58b02bf37b6d","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Neural rgb-d surface reconstruction","work_id":"6d621e0a-3d83-421a-bde8-d6e715851bf6","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Token merging: Your vit but faster","work_id":"79e860e9-4223-4e36-8b25-31f593452d95","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"A naturalistic open source movie for optical flow evaluation","work_id":"3f58f385-64bd-4f4e-beb9-0981c26077c8","year":2012}],"snapshot_sha256":"e3c5dfec023fd80a54cb702ecaff3818f95acf34f01bd99625dbb2f32d63efd8"},"source":{"id":"2605.14315","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T02:15:57.914950Z","id":"ce52b30b-5e1b-42b6-b903-49bbc5fc6268","model_set":{"reader":"grok-4.3"},"one_line_summary":"TurboVGGT uses adaptive sparse global attention with varying sparsity levels across frames and layers plus frame attention to enable faster multi-view 3D reconstruction while keeping competitive quality versus prior state-of-the-art methods.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"TurboVGGT speeds multi-view 3D reconstruction by learning varying sparsity in attention across frames and layers.","strongest_claim":"TurboVGGT achieves fast multi-view reconstruction while maintaining competitive reconstruction quality compared with state-of-the-art methods.","weakest_assumption":"The assumption that adaptively learning representative tokens with varying sparsity levels across frames, layers, and structurally informative regions will reliably capture global relationships without losing critical geometric details in diverse real-world scenes."}},"verdict_id":"ce52b30b-5e1b-42b6-b903-49bbc5fc6268"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:5a8af6fa8dd9ea476269bdd589408985441e8d7546bc6970b773740c6c1f162a","target":"record","created_at":"2026-05-17T23:39:09Z","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":"831cda8484bd628f8a5135b3999062c1115841ad209086668e028b944115fa15","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-14T03:24:09Z","title_canon_sha256":"28ec58e343c854ab448e2b5a731c91a40049bf1c163053032dd5760388681c72"},"schema_version":"1.0","source":{"id":"2605.14315","kind":"arxiv","version":1}},"canonical_sha256":"52f9db328c4eade098423a5ef2557ba74940707614afe1352b13c3b3307e9f89","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"52f9db328c4eade098423a5ef2557ba74940707614afe1352b13c3b3307e9f89","first_computed_at":"2026-05-17T23:39:09.905428Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:09.905428Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"S6/D8qts75XDaoZSB0+9Z+Jo1Tw8cAtm/geEAcVLxh2qtOZWSeIUsUx1c52bJqFeMNLbeCM1t2d3+2WVnxg2Dg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:09.906037Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.14315","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5a8af6fa8dd9ea476269bdd589408985441e8d7546bc6970b773740c6c1f162a","sha256:6eaa0846f616c9d8a8297dda0876be1fc6954b22f0d79ce736d0e5441bd5252b"],"state_sha256":"217cf36f81829ce5778a88b57c688225b1fd200d12095d8dd982c2608baaf852"}