{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:W7EVYY73AJQPRBOQGAB6K7EDNE","short_pith_number":"pith:W7EVYY73","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"},"canonical_sha256":"b7c95c63fb0260f885d03003e57c83690a56bdacde9bc816fbfbd059b4ca1e07","source":{"kind":"arxiv","id":"2509.26645","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.26645","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"arxiv_version","alias_value":"2509.26645v4","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.26645","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"pith_short_12","alias_value":"W7EVYY73AJQP","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"W7EVYY73AJQPRBOQ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"W7EVYY73","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:W7EVYY73AJQPRBOQGAB6K7EDNE","target":"record","payload":{"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"},"canonical_sha256":"b7c95c63fb0260f885d03003e57c83690a56bdacde9bc816fbfbd059b4ca1e07","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"},"source_kind":"arxiv","source_id":"2509.26645","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DMK5Ws/ve8V1lRVr4LokCfog6uZWOYFJPXCTUJ7pxTTQ5VUfZW231r+xqQsfgmHM8Rj5gWwZgv4rwk4I8pX1Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-19T23:27:54.176156Z"},"content_sha256":"7e0f2ee534910af9f264016cfe7a37441ab638fed62513ced8195f02d550036c","schema_version":"1.0","event_id":"sha256:7e0f2ee534910af9f264016cfe7a37441ab638fed62513ced8195f02d550036c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:W7EVYY73AJQPRBOQGAB6K7EDNE","target":"graph","payload":{"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"},"verdict_id":"eca97549-12f1-47ef-b8be-f9650a31484c"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8IeN1Z3bxSELhtWsF76HH1Da6zRn49+Rwj26tI+ifdlE/j+ggj5lOenDyBGzRQ0x6rSFJdjOK5rKX+MYjAlJBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-19T23:27:54.177194Z"},"content_sha256":"5d31dcd7970b7076c3958a6aa86da126a3d747ac5eadc25d78912efcb40b1a4c","schema_version":"1.0","event_id":"sha256:5d31dcd7970b7076c3958a6aa86da126a3d747ac5eadc25d78912efcb40b1a4c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/bundle.json","state_url":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/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":"2026-05-19T23:27:54Z","links":{"resolver":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE","bundle":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/bundle.json","state":"https://pith.science/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/W7EVYY73AJQPRBOQGAB6K7EDNE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:W7EVYY73AJQPRBOQGAB6K7EDNE","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":"faf78cda85d0821e0ca3b7bc112f727b3a7f0ffb3ce8164936548f6b91e0d2d5","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-09-30T17:59:51Z","title_canon_sha256":"a769ed00c8dcbbcee71e9d931c89007e9966939155db4db9d109d16b780217c0"},"schema_version":"1.0","source":{"id":"2509.26645","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.26645","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"arxiv_version","alias_value":"2509.26645v4","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.26645","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"pith_short_12","alias_value":"W7EVYY73AJQP","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"W7EVYY73AJQPRBOQ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"W7EVYY73","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:5d31dcd7970b7076c3958a6aa86da126a3d747ac5eadc25d78912efcb40b1a4c","target":"graph","created_at":"2026-05-17T23:38:14Z","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":"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."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","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."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","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."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","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."}],"snapshot_sha256":"7c70ed132bc01dfcaafee50c054a19a61e361953e5e3792002f4c25e4cb22fa6"},"formal_canon":{"evidence_count":1,"snapshot_sha256":"a204f5ebae18cf3a9b3b7135dc56cb4c5a24e002eb0a68948a28b7e507b242d8"},"paper":{"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","authors_text":"Andreas Geiger, Anpei Chen, Xingyu Chen, Yue Chen, Yuliang Xiu","cross_cats":[],"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.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-09-30T17:59:51Z","title":"TTT3R: 3D Reconstruction as Test-Time Training"},"references":{"count":112,"internal_anchors":18,"resolved_work":112,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Bundle adjustment in the large","work_id":"98511d9d-34e5-46fc-b399-cfed80b316d7","year":2010},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Building rome in a day.ACM Communications, 2011","work_id":"66df7549-0359-4b87-b4aa-d40cc0986570","year":2011},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Cross-view completion models are zero-shot correspondence estimators","work_id":"1b5feeac-d1ba-40fc-b2af-03fce6a3e07e","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Speeded-up robust features (surf).Computer vision and image understanding, 2008","work_id":"3620411b-df89-473c-ba6d-5c6184b527bb","year":2008},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"xlstm: Extended long short-term memory","work_id":"44e150f9-d858-4df1-9dc2-1cae87bc20fc","year":2024}],"snapshot_sha256":"e0089e70acc92b31e4aaafba12806174d7da72bf50922815eebee91a3e593664"},"source":{"id":"2509.26645","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-17T06:35:59.893810Z","id":"eca97549-12f1-47ef-b8be-f9650a31484c","model_set":{"reader":"grok-4.3"},"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","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.","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.","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."}},"verdict_id":"eca97549-12f1-47ef-b8be-f9650a31484c"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:7e0f2ee534910af9f264016cfe7a37441ab638fed62513ced8195f02d550036c","target":"record","created_at":"2026-05-17T23:38:14Z","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":"faf78cda85d0821e0ca3b7bc112f727b3a7f0ffb3ce8164936548f6b91e0d2d5","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-09-30T17:59:51Z","title_canon_sha256":"a769ed00c8dcbbcee71e9d931c89007e9966939155db4db9d109d16b780217c0"},"schema_version":"1.0","source":{"id":"2509.26645","kind":"arxiv","version":4}},"canonical_sha256":"b7c95c63fb0260f885d03003e57c83690a56bdacde9bc816fbfbd059b4ca1e07","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b7c95c63fb0260f885d03003e57c83690a56bdacde9bc816fbfbd059b4ca1e07","first_computed_at":"2026-05-17T23:38:14.802949Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:14.802949Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"UVbrpmBpf2l0FFIBQHxbu/GjHGY6CSur5TvJcDSg5w6JqjHpEw0jieToDeLfbYD0kKcpS3vs/GA+YkWgFIq/AA==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:14.803626Z","signed_message":"canonical_sha256_bytes"},"source_id":"2509.26645","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7e0f2ee534910af9f264016cfe7a37441ab638fed62513ced8195f02d550036c","sha256:5d31dcd7970b7076c3958a6aa86da126a3d747ac5eadc25d78912efcb40b1a4c"],"state_sha256":"152367a38d4a9f6c0a5ba33d6911b3564bcc2cd3eb9f349f21bf408de13674ff"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vmzXB9RP3/EjkRUvekoRi55Vb1V026i0xLPcbx8t2955bSuivYBau5n6ZvowOk8wYq/glq4O4k6pkGJyd9cSDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-19T23:27:54.181636Z","bundle_sha256":"b69f55115fd705b7e3ee4c33264ff19d01d164247522be29b03a7711c3fd4444"}}