{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:4T3TOVISQ2Z2DPIVZ3AWRO5RK4","short_pith_number":"pith:4T3TOVIS","schema_version":"1.0","canonical_sha256":"e4f737551286b3a1bd15cec168bbb1573f9cf68dbd5e682fe6d5746ce9834276","source":{"kind":"arxiv","id":"2203.15081","version":5},"attestation_state":"computed","paper":{"title":"Word Discovery in Visually Grounded, Self-Supervised Speech Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.SD"],"primary_cat":"eess.AS","authors_text":"David Harwath, Puyuan Peng","submitted_at":"2022-03-28T20:41:17Z","abstract_excerpt":"We present a method for visually-grounded spoken term discovery. After training either a HuBERT or wav2vec2.0 model to associate spoken captions with natural images, we show that powerful word segmentation and clustering capability emerges within the model's self-attention heads. Our experiments reveal that this ability is not present to nearly the same extent in the base HuBERT and wav2vec2.0 models, suggesting that the visual grounding task is a crucial component of the word discovery capability we observe. We also evaluate our method on the Buckeye word segmentation and ZeroSpeech spoken te"},"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":"2203.15081","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2022-03-28T20:41:17Z","cross_cats_sorted":["cs.AI","cs.CL","cs.SD"],"title_canon_sha256":"9b8d0476933932e8a897bd41ab2026771bbd93b0036332b5f7f85b92b98223c7","abstract_canon_sha256":"36478dd9499109834de5e0cd798d80941a295c5422b2fe694e5a98b61b0d36a4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:21:46.402985Z","signature_b64":"Ro2MUmw4xU0JBmtuPu0w3M/153y2Tt4KTaec8L8UHx3DRhem7lf96Nq/Qsv9LmFJkqxj4txsGvBxRS827XxhCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e4f737551286b3a1bd15cec168bbb1573f9cf68dbd5e682fe6d5746ce9834276","last_reissued_at":"2026-07-05T06:21:46.402569Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:21:46.402569Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Word Discovery in Visually Grounded, Self-Supervised Speech Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.SD"],"primary_cat":"eess.AS","authors_text":"David Harwath, Puyuan Peng","submitted_at":"2022-03-28T20:41:17Z","abstract_excerpt":"We present a method for visually-grounded spoken term discovery. After training either a HuBERT or wav2vec2.0 model to associate spoken captions with natural images, we show that powerful word segmentation and clustering capability emerges within the model's self-attention heads. Our experiments reveal that this ability is not present to nearly the same extent in the base HuBERT and wav2vec2.0 models, suggesting that the visual grounding task is a crucial component of the word discovery capability we observe. We also evaluate our method on the Buckeye word segmentation and ZeroSpeech spoken te"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.15081","kind":"arxiv","version":5},"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/2203.15081/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":"2203.15081","created_at":"2026-07-05T06:21:46.402625+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.15081v5","created_at":"2026-07-05T06:21:46.402625+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.15081","created_at":"2026-07-05T06:21:46.402625+00:00"},{"alias_kind":"pith_short_12","alias_value":"4T3TOVISQ2Z2","created_at":"2026-07-05T06:21:46.402625+00:00"},{"alias_kind":"pith_short_16","alias_value":"4T3TOVISQ2Z2DPIV","created_at":"2026-07-05T06:21:46.402625+00:00"},{"alias_kind":"pith_short_8","alias_value":"4T3TOVIS","created_at":"2026-07-05T06:21:46.402625+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.01381","citing_title":"A framework for analyzing concept representations in neural models","ref_index":255,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4","json":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4.json","graph_json":"https://pith.science/api/pith-number/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/graph.json","events_json":"https://pith.science/api/pith-number/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/events.json","paper":"https://pith.science/paper/4T3TOVIS"},"agent_actions":{"view_html":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4","download_json":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4.json","view_paper":"https://pith.science/paper/4T3TOVIS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.15081&json=true","fetch_graph":"https://pith.science/api/pith-number/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/graph.json","fetch_events":"https://pith.science/api/pith-number/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/action/storage_attestation","attest_author":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/action/author_attestation","sign_citation":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/action/citation_signature","submit_replication":"https://pith.science/pith/4T3TOVISQ2Z2DPIVZ3AWRO5RK4/action/replication_record"}},"created_at":"2026-07-05T06:21:46.402625+00:00","updated_at":"2026-07-05T06:21:46.402625+00:00"}