{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:KGO5X5TIAS7PFY25TKMFWZPJU2","short_pith_number":"pith:KGO5X5TI","schema_version":"1.0","canonical_sha256":"519ddbf66804bef2e35d9a985b65e9a696ee9c1f1897a53ac6a3fc5e37d165bb","source":{"kind":"arxiv","id":"2202.03555","version":3},"attestation_state":"computed","paper":{"title":"data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alexei Baevski, Arun Babu, Jiatao Gu, Michael Auli, Qiantong Xu, Wei-Ning Hsu","submitted_at":"2022-02-07T22:52:11Z","abstract_excerpt":"While the general idea of self-supervised learning is identical across modalities, the actual algorithms and objectives differ widely because they were developed with a single modality in mind. To get us closer to general self-supervised learning, we present data2vec, a framework that uses the same learning method for either speech, NLP or computer vision. The core idea is to predict latent representations of the full input data based on a masked view of the input in a self-distillation setup using a standard Transformer architecture. Instead of predicting modality-specific targets such as wor"},"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":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2202.03555","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2022-02-07T22:52:11Z","cross_cats_sorted":[],"title_canon_sha256":"215b9661dab5df68adca48230fad995fb42d8a78c1ffbde1c6d3fbfc1939235d","abstract_canon_sha256":"0c49dc2c088fa0af5c2db5a45354779b8fc099180cf9b1a9b1d8df5fa30cc71c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:10:36.503400Z","signature_b64":"NGPDIzWSypQtu/3V0Uz+gjSTB1OuRuG/AXDOPEr2TxadF/3eM92F09KfiLbIBi72mFP+lJzysXMyokyKpzSMCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"519ddbf66804bef2e35d9a985b65e9a696ee9c1f1897a53ac6a3fc5e37d165bb","last_reissued_at":"2026-07-05T05:10:36.502894Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:10:36.502894Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alexei Baevski, Arun Babu, Jiatao Gu, Michael Auli, Qiantong Xu, Wei-Ning Hsu","submitted_at":"2022-02-07T22:52:11Z","abstract_excerpt":"While the general idea of self-supervised learning is identical across modalities, the actual algorithms and objectives differ widely because they were developed with a single modality in mind. To get us closer to general self-supervised learning, we present data2vec, a framework that uses the same learning method for either speech, NLP or computer vision. The core idea is to predict latent representations of the full input data based on a masked view of the input in a self-distillation setup using a standard Transformer architecture. Instead of predicting modality-specific targets such as wor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.03555","kind":"arxiv","version":3},"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/2202.03555/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2202.03555","created_at":"2026-07-05T05:10:36.502953+00:00"},{"alias_kind":"arxiv_version","alias_value":"2202.03555v3","created_at":"2026-07-05T05:10:36.502953+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.03555","created_at":"2026-07-05T05:10:36.502953+00:00"},{"alias_kind":"pith_short_12","alias_value":"KGO5X5TIAS7P","created_at":"2026-07-05T05:10:36.502953+00:00"},{"alias_kind":"pith_short_16","alias_value":"KGO5X5TIAS7PFY25","created_at":"2026-07-05T05:10:36.502953+00:00"},{"alias_kind":"pith_short_8","alias_value":"KGO5X5TI","created_at":"2026-07-05T05:10:36.502953+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":15,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2607.06629","citing_title":"STST-JEPA: Shallow-Target Spatio-Temporal Joint Embedding Prediction Architecture For EEG Self-Supervised Learning","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2607.07756","citing_title":"The Importance of Encoder Choice:A Tabular-Image Study","ref_index":253,"is_internal_anchor":true},{"citing_arxiv_id":"2606.25225","citing_title":"MJEPA: A Simple and Scalable Joint-Embedding Predictive Architecture for Audio-Visual Learning","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21052","citing_title":"Backdoor Attacks on Speech Emotion Recognition via TTS-Generated Poisoning","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.24910","citing_title":"End-to-End Voice Intent Recognition for Spontaneous Human-Drone Interaction with Naive Users","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2606.11190","citing_title":"When to Align, When to Predict: A Phase Diagram for Multimodal Learning","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09335","citing_title":"Factors affecting ASR performance: A study using state of the art ASR models in Indic Languages","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00052","citing_title":"AGE: Adaptive-masking for Graph Embedding in Graph Retrieval-Augmented Generation","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01145","citing_title":"Hierarchical Self-Supervised Representation Learning Framework for Multivariate Time Series Grounded in ECG Analysis","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.11190","citing_title":"When to Align, When to Predict: A Phase Diagram for Multimodal Learning","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11654","citing_title":"Weather-Robust Cross-View Geo-Localization via Prototype-Based Semantic Part Discovery","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15394","citing_title":"Representation Without Reward: A JEPA Audit for LLM Fine-Tuning","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2603.01482","citing_title":"A SUPERB-Style Benchmark of Self-Supervised Speech Models for Audio Deepfake Detection","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11654","citing_title":"Weather-Robust Cross-View Geo-Localization via Prototype-Based Semantic Part Discovery","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2404.08471","citing_title":"Revisiting Feature Prediction for Learning Visual Representations from Video","ref_index":223,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2","json":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2.json","graph_json":"https://pith.science/api/pith-number/KGO5X5TIAS7PFY25TKMFWZPJU2/graph.json","events_json":"https://pith.science/api/pith-number/KGO5X5TIAS7PFY25TKMFWZPJU2/events.json","paper":"https://pith.science/paper/KGO5X5TI"},"agent_actions":{"view_html":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2","download_json":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2.json","view_paper":"https://pith.science/paper/KGO5X5TI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2202.03555&json=true","fetch_graph":"https://pith.science/api/pith-number/KGO5X5TIAS7PFY25TKMFWZPJU2/graph.json","fetch_events":"https://pith.science/api/pith-number/KGO5X5TIAS7PFY25TKMFWZPJU2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2/action/storage_attestation","attest_author":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2/action/author_attestation","sign_citation":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2/action/citation_signature","submit_replication":"https://pith.science/pith/KGO5X5TIAS7PFY25TKMFWZPJU2/action/replication_record"}},"created_at":"2026-07-05T05:10:36.502953+00:00","updated_at":"2026-07-05T05:10:36.502953+00:00"}