{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6EHDR6IZELD7BOYYKI64LROOOU","short_pith_number":"pith:6EHDR6IZ","schema_version":"1.0","canonical_sha256":"f10e38f91922c7f0bb18523dc5c5ce7510545c21e1259786aa0595bec76d5a79","source":{"kind":"arxiv","id":"2605.16775","version":1},"attestation_state":"computed","paper":{"title":"VolTA-3D: Self-Supervised Learning for Brain MRI using 3D Volumetric Token Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"VolTA-3D aligns global class-style tokens and local patch tokens in a student-teacher setup to learn transferable 3D representations from unlabeled brain MRI.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Abhijeet Parida, Amy Makawana, Julia Ive, Marius George Linguraru, Syed Muhammad Anwar","submitted_at":"2026-05-16T03:09:25Z","abstract_excerpt":"Self-supervised learning (SSL) has advanced medical image analysis be enabling learning form large unlabelled data. However, in brain magnetic resonance imaging (MRI), most 3D models remain specialized for either segmentation of classification, limiting their ability to generalize across datasets, imaging protocols,, and downstream tasks. This lack of transferability constrains the clinical utility of 3D MRI models, despite the availability of unlabeled volumetric data. We present Volta-3D, a self-supervised 3D Vision Transformer framework designed to learn transferable volumetric representati"},"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":"2605.16775","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-16T03:09:25Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"317aeb0c0fa03a8db06f88f2b6bd5bf78b417c759c4e8b06b14113201728b8eb","abstract_canon_sha256":"732e438c0cdce275f0c84d82977a54a1a1c1da7fdef5231000ccb29e7321e1f8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:21.343635Z","signature_b64":"+F1Dt7CEJx2r1n3DBWmXaGtI+os+AkWI3N2UcqbXzI6IjjcVPqjuCdWca0UQfzPJqibLuGaFq3ETgyxv3MFeAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f10e38f91922c7f0bb18523dc5c5ce7510545c21e1259786aa0595bec76d5a79","last_reissued_at":"2026-05-20T00:03:21.342709Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:21.342709Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VolTA-3D: Self-Supervised Learning for Brain MRI using 3D Volumetric Token Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"VolTA-3D aligns global class-style tokens and local patch tokens in a student-teacher setup to learn transferable 3D representations from unlabeled brain MRI.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Abhijeet Parida, Amy Makawana, Julia Ive, Marius George Linguraru, Syed Muhammad Anwar","submitted_at":"2026-05-16T03:09:25Z","abstract_excerpt":"Self-supervised learning (SSL) has advanced medical image analysis be enabling learning form large unlabelled data. However, in brain magnetic resonance imaging (MRI), most 3D models remain specialized for either segmentation of classification, limiting their ability to generalize across datasets, imaging protocols,, and downstream tasks. This lack of transferability constrains the clinical utility of 3D MRI models, despite the availability of unlabeled volumetric data. We present Volta-3D, a self-supervised 3D Vision Transformer framework designed to learn transferable volumetric representati"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Hence jointly enforcing global semantic consistency and local structural learning during pretraining enables broader concept learning from unlabeled brain MRI data. Overall VolTA-3D supports effective multi-task downstream performance with task-specific pretraining, a step towards generalizable and clinically viable 3D models.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The premise that the limited semantic diversity and subtle anatomical characteristics of brain MRI specifically challenge existing SSL approaches and that the proposed global-local token alignment within a student-teacher paradigm will overcome these challenges to produce improved transferability and robustness under domain shift.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"VolTA-3D learns transferable 3D representations from unlabeled brain MRI by jointly aligning global and local tokens in a self-supervised student-teacher framework.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"VolTA-3D aligns global class-style tokens and local patch tokens in a student-teacher setup to learn transferable 3D representations from unlabeled brain MRI.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"930b1a965134289d7c258633e5b66a1b20d71812925a55e369326c7ae4e70a04"},"source":{"id":"2605.16775","kind":"arxiv","version":1},"verdict":{"id":"ebba34ea-d8b3-4785-96e6-738978dff9b7","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T21:44:19.540369Z","strongest_claim":"Hence jointly enforcing global semantic consistency and local structural learning during pretraining enables broader concept learning from unlabeled brain MRI data. Overall VolTA-3D supports effective multi-task downstream performance with task-specific pretraining, a step towards generalizable and clinically viable 3D models.","one_line_summary":"VolTA-3D learns transferable 3D representations from unlabeled brain MRI by jointly aligning global and local tokens in a self-supervised student-teacher framework.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The premise that the limited semantic diversity and subtle anatomical characteristics of brain MRI specifically challenge existing SSL approaches and that the proposed global-local token alignment within a student-teacher paradigm will overcome these challenges to produce improved transferability and robustness under domain shift.","pith_extraction_headline":"VolTA-3D aligns global class-style tokens and local patch tokens in a student-teacher setup to learn transferable 3D representations from unlabeled brain MRI."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.16775/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T22:01:19.721273Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T21:50:55.310765Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T19:01:56.307134Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.441765Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"1c8f3429355662c5537b1a3e8475ea2ec6ef7ccd60bea590277e0a46ba9c9939"},"references":{"count":23,"sample":[{"doi":"","year":2021,"title":"Workload of diagnostic radiologists in the foreseeable future based on recent scientific advances: growth expectations and role of artificial intelligence,","work_id":"cfcd1300-8a51-425e-9104-87686f2b89b3","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2015,"title":"Mri seg- mentation of the human brain: Challenges, methods, and applications,","work_id":"f5ebfd44-3068-43ef-8285-caea408ec7f4","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Building a general simclr self-supervised foundation model across neurological diseases to advance 3d brain mri diagnoses,","work_id":"19b36503-1ea0-44c4-a20a-47a6d78daa98","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Domain adaptation for medical image analysis: A survey,","work_id":"a67a9a37-78ec-4a0d-a733-134862185dc6","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Comparing 3d, 2.5 d, and 2d approaches to brain image auto-segmentation,","work_id":"d3bc9332-3c03-43f5-9ca5-fe6b7fc63841","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":23,"snapshot_sha256":"57dfcfdaa5d0ab489705344ca5efa45c2f5e91af0b0ff24a58cfd17a2cc2d2b7","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"949480d3545b0e8053d84ac89362707dcdf186e7bc08eb9dda202c74d1e54d61"},"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":"2605.16775","created_at":"2026-05-20T00:03:21.342858+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.16775v1","created_at":"2026-05-20T00:03:21.342858+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16775","created_at":"2026-05-20T00:03:21.342858+00:00"},{"alias_kind":"pith_short_12","alias_value":"6EHDR6IZELD7","created_at":"2026-05-20T00:03:21.342858+00:00"},{"alias_kind":"pith_short_16","alias_value":"6EHDR6IZELD7BOYY","created_at":"2026-05-20T00:03:21.342858+00:00"},{"alias_kind":"pith_short_8","alias_value":"6EHDR6IZ","created_at":"2026-05-20T00:03:21.342858+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU","json":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU.json","graph_json":"https://pith.science/api/pith-number/6EHDR6IZELD7BOYYKI64LROOOU/graph.json","events_json":"https://pith.science/api/pith-number/6EHDR6IZELD7BOYYKI64LROOOU/events.json","paper":"https://pith.science/paper/6EHDR6IZ"},"agent_actions":{"view_html":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU","download_json":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU.json","view_paper":"https://pith.science/paper/6EHDR6IZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.16775&json=true","fetch_graph":"https://pith.science/api/pith-number/6EHDR6IZELD7BOYYKI64LROOOU/graph.json","fetch_events":"https://pith.science/api/pith-number/6EHDR6IZELD7BOYYKI64LROOOU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU/action/storage_attestation","attest_author":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU/action/author_attestation","sign_citation":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU/action/citation_signature","submit_replication":"https://pith.science/pith/6EHDR6IZELD7BOYYKI64LROOOU/action/replication_record"}},"created_at":"2026-05-20T00:03:21.342858+00:00","updated_at":"2026-05-20T00:03:21.342858+00:00"}