{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:W7EVYY73AJQPRBOQGAB6K7EDNE","short_pith_number":"pith:W7EVYY73","schema_version":"1.0","canonical_sha256":"b7c95c63fb0260f885d03003e57c83690a56bdacde9bc816fbfbd059b4ca1e07","source":{"kind":"arxiv","id":"2509.26645","version":4},"attestation_state":"computed","paper":{"title":"TTT3R: 3D Reconstruction as Test-Time Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Framing 3D reconstruction as test-time training yields a closed-form learning rate from alignment confidence that doubles global pose accuracy on long sequences.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Andreas Geiger, Anpei Chen, Xingyu Chen, Yue Chen, Yuliang Xiu","submitted_at":"2025-09-30T17:59:51Z","abstract_excerpt":"Modern Recurrent Neural Networks have become a competitive architecture for 3D reconstruction due to their linear-time complexity. However, their performance degrades significantly when applied beyond the training context length, revealing limited length generalization. In this work, we revisit the 3D reconstruction foundation models from a Test-Time Training perspective, framing their designs as an online learning problem. Building on this perspective, we leverage the alignment confidence between the memory state and incoming observations to derive a closed-form learning rate for memory updat"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2509.26645","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-09-30T17:59:51Z","cross_cats_sorted":[],"title_canon_sha256":"a769ed00c8dcbbcee71e9d931c89007e9966939155db4db9d109d16b780217c0","abstract_canon_sha256":"faf78cda85d0821e0ca3b7bc112f727b3a7f0ffb3ce8164936548f6b91e0d2d5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:14.803626Z","signature_b64":"UVbrpmBpf2l0FFIBQHxbu/GjHGY6CSur5TvJcDSg5w6JqjHpEw0jieToDeLfbYD0kKcpS3vs/GA+YkWgFIq/AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b7c95c63fb0260f885d03003e57c83690a56bdacde9bc816fbfbd059b4ca1e07","last_reissued_at":"2026-05-17T23:38:14.802949Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:14.802949Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TTT3R: 3D Reconstruction as Test-Time Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Framing 3D reconstruction as test-time training yields a closed-form learning rate from alignment confidence that doubles global pose accuracy on long sequences.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Andreas Geiger, Anpei Chen, Xingyu Chen, Yue Chen, Yuliang Xiu","submitted_at":"2025-09-30T17:59:51Z","abstract_excerpt":"Modern Recurrent Neural Networks have become a competitive architecture for 3D reconstruction due to their linear-time complexity. However, their performance degrades significantly when applied beyond the training context length, revealing limited length generalization. In this work, we revisit the 3D reconstruction foundation models from a Test-Time Training perspective, framing their designs as an online learning problem. Building on this perspective, we leverage the alignment confidence between the memory state and incoming observations to derive a closed-form learning rate for memory updat"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"This training-free intervention, termed TTT3R, substantially improves length generalization, achieving a 2× improvement in global pose estimation over baselines, while operating at 20 FPS with just 6 GB of GPU memory to process thousands of images.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That alignment confidence between the memory state and incoming observations can be computed reliably and directly yields a closed-form learning rate that correctly balances retention of history with adaptation to new data without introducing instability or bias.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TTT3R derives a closed-form learning rate from memory-observation alignment confidence to boost length generalization in RNN-based 3D reconstruction by 2x in global pose estimation.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Framing 3D reconstruction as test-time training yields a closed-form learning rate from alignment confidence that doubles global pose accuracy on long sequences.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"7c70ed132bc01dfcaafee50c054a19a61e361953e5e3792002f4c25e4cb22fa6"},"source":{"id":"2509.26645","kind":"arxiv","version":4},"verdict":{"id":"eca97549-12f1-47ef-b8be-f9650a31484c","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T06:35:59.893810Z","strongest_claim":"This training-free intervention, termed TTT3R, substantially improves length generalization, achieving a 2× improvement in global pose estimation over baselines, while operating at 20 FPS with just 6 GB of GPU memory to process thousands of images.","one_line_summary":"TTT3R derives a closed-form learning rate from memory-observation alignment confidence to boost length generalization in RNN-based 3D reconstruction by 2x in global pose estimation.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That alignment confidence between the memory state and incoming observations can be computed reliably and directly yields a closed-form learning rate that correctly balances retention of history with adaptation to new data without introducing instability or bias.","pith_extraction_headline":"Framing 3D reconstruction as test-time training yields a closed-form learning rate from alignment confidence that doubles global pose accuracy on long sequences."},"references":{"count":112,"sample":[{"doi":"","year":2010,"title":"Bundle adjustment in the large","work_id":"98511d9d-34e5-46fc-b399-cfed80b316d7","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2011,"title":"Building rome in a day.ACM Communications, 2011","work_id":"66df7549-0359-4b87-b4aa-d40cc0986570","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Cross-view completion models are zero-shot correspondence estimators","work_id":"1b5feeac-d1ba-40fc-b2af-03fce6a3e07e","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2008,"title":"Speeded-up robust features (surf).Computer vision and image understanding, 2008","work_id":"3620411b-df89-473c-ba6d-5c6184b527bb","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"xlstm: Extended long short-term memory","work_id":"44e150f9-d858-4df1-9dc2-1cae87bc20fc","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":112,"snapshot_sha256":"e0089e70acc92b31e4aaafba12806174d7da72bf50922815eebee91a3e593664","internal_anchors":18},"formal_canon":{"evidence_count":1,"snapshot_sha256":"a204f5ebae18cf3a9b3b7135dc56cb4c5a24e002eb0a68948a28b7e507b242d8"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2509.26645","created_at":"2026-05-17T23:38:14.803056+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.26645v4","created_at":"2026-05-17T23:38:14.803056+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.26645","created_at":"2026-05-17T23:38:14.803056+00:00"},{"alias_kind":"pith_short_12","alias_value":"W7EVYY73AJQP","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"W7EVYY73AJQPRBOQ","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"W7EVYY73","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":19,"internal_anchor_count":19,"sample":[{"citing_arxiv_id":"2605.16981","citing_title":"Rethinking the State Update Gate for Long-Sequence Recurrent 3D Reconstruction","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2512.01643","citing_title":"ViT$^3$: Unlocking Test-Time Training in Vision","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2512.15577","citing_title":"MoonSeg3R: Monocular Online Zero-Shot Segment Anything in 3D with Reconstructive Foundation Priors","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2603.04385","citing_title":"ZipMap: Linear-Time Stateful 3D Reconstruction via Test-Time Training","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2603.05959","citing_title":"OVGGT: O(1) Constant-Cost Streaming Visual Geometry Transformer","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2603.07690","citing_title":"FrameVGGT: Geometry-Aligned Frame-Level Memory for Bounded Streaming VGGT","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2604.03878","citing_title":"Learning 3D Reconstruction with Priors in Test Time","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2605.05749","citing_title":"Ray-Aware Pointer Memory with Adaptive Updates for Streaming 3D Reconstruction","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09644","citing_title":"Attention Itself Could Retrieve.RetrieveVGGT: Training-Free Long Context Streaming 3D Reconstruction via Query-Key Similarity Retrieval","ref_index":26,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06270","citing_title":"Spark3R: Asymmetric Token Reduction Makes Fast Feed-Forward 3D Reconstruction","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2605.05749","citing_title":"Ray-Aware Pointer Memory with Adaptive Updates for Streaming 3D Reconstruction","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2604.08542","citing_title":"Scal3R: Scalable Test-Time Training for Large-Scale 3D Reconstruction","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2604.05351","citing_title":"AnyImageNav: Any-View Geometry for Precise Last-Meter Image-Goal Navigation","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2604.13476","citing_title":"RobotPan: A 360$^\\circ$ Surround-View Robotic Vision System for Embodied Perception","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2604.14025","citing_title":"Feed-Forward 3D Scene Modeling: A Problem-Driven Perspective","ref_index":109,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15239","citing_title":"TokenGS: Decoupling 3D Gaussian Prediction from Pixels with Learnable Tokens","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15237","citing_title":"StreamCacheVGGT: Streaming Visual Geometry Transformers with Robust Scoring and Hybrid Cache Compression","ref_index":37,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15284","citing_title":"GlobalSplat: Efficient Feed-Forward 3D Gaussian Splatting via Global Scene Tokens","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.02772","citing_title":"Linearizing Vision Transformer with Test-Time Training","ref_index":4,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":1,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE","json":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE.json","graph_json":"https://pith.science/api/pith-number/W7EVYY73AJQPRBOQGAB6K7EDNE/graph.json","events_json":"https://pith.science/api/pith-number/W7EVYY73AJQPRBOQGAB6K7EDNE/events.json","paper":"https://pith.science/paper/W7EVYY73"},"agent_actions":{"view_html":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE","download_json":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE.json","view_paper":"https://pith.science/paper/W7EVYY73","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.26645&json=true","fetch_graph":"https://pith.science/api/pith-number/W7EVYY73AJQPRBOQGAB6K7EDNE/graph.json","fetch_events":"https://pith.science/api/pith-number/W7EVYY73AJQPRBOQGAB6K7EDNE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/action/storage_attestation","attest_author":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/action/author_attestation","sign_citation":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/action/citation_signature","submit_replication":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/action/replication_record"}},"created_at":"2026-05-17T23:38:14.803056+00:00","updated_at":"2026-05-17T23:38:14.803056+00:00"}