{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4XGPFRNHLSQROELL2LKQMK3QFL","short_pith_number":"pith:4XGPFRNH","schema_version":"1.0","canonical_sha256":"e5ccf2c5a75ca117116bd2d5062b702ad01a45b92642380ff0cd6edc9dd8a959","source":{"kind":"arxiv","id":"2506.17690","version":1},"attestation_state":"computed","paper":{"title":"Low-resource keyword spotting using contrastively trained transformer acoustic word embeddings","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"eess.AS","authors_text":"Christiaan Jacobs, Julian Herreilers, Thomas Niesler","submitted_at":"2025-06-21T11:47:05Z","abstract_excerpt":"We introduce a new approach, the ContrastiveTransformer, that produces acoustic word embeddings (AWEs) for the purpose of very low-resource keyword spotting. The ContrastiveTransformer, an encoder-only model, directly optimises the embedding space using normalised temperature-scaled cross entropy (NT-Xent) loss. We use this model to perform keyword spotting for radio broadcasts in Luganda and Bambara, the latter a severely under-resourced language. We compare our model to various existing AWE approaches, including those constructed from large pre-trained self-supervised models, a recurrent enc"},"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":"2506.17690","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2025-06-21T11:47:05Z","cross_cats_sorted":[],"title_canon_sha256":"df03602215ad4f1906fa8bcbc606b309eb51a3cb02278dca86f5b9a9796e6003","abstract_canon_sha256":"33405bc37d14c1894a8f1340fd77d9e4fee6f0f09b2beeb044413c4cf5343748"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:25:21.908987Z","signature_b64":"iqHQ43JqFRLd4r3Dy7GmUJB2/rJj8cPKcNwjHQ0ElsacyNVBCQP5D4X8BbDRAdPzbavAQvYuq91OVFWZijMaBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e5ccf2c5a75ca117116bd2d5062b702ad01a45b92642380ff0cd6edc9dd8a959","last_reissued_at":"2026-07-05T11:25:21.908447Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:25:21.908447Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Low-resource keyword spotting using contrastively trained transformer acoustic word embeddings","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"eess.AS","authors_text":"Christiaan Jacobs, Julian Herreilers, Thomas Niesler","submitted_at":"2025-06-21T11:47:05Z","abstract_excerpt":"We introduce a new approach, the ContrastiveTransformer, that produces acoustic word embeddings (AWEs) for the purpose of very low-resource keyword spotting. The ContrastiveTransformer, an encoder-only model, directly optimises the embedding space using normalised temperature-scaled cross entropy (NT-Xent) loss. We use this model to perform keyword spotting for radio broadcasts in Luganda and Bambara, the latter a severely under-resourced language. We compare our model to various existing AWE approaches, including those constructed from large pre-trained self-supervised models, a recurrent enc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.17690","kind":"arxiv","version":1},"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/2506.17690/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":"2506.17690","created_at":"2026-07-05T11:25:21.908507+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.17690v1","created_at":"2026-07-05T11:25:21.908507+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.17690","created_at":"2026-07-05T11:25:21.908507+00:00"},{"alias_kind":"pith_short_12","alias_value":"4XGPFRNHLSQR","created_at":"2026-07-05T11:25:21.908507+00:00"},{"alias_kind":"pith_short_16","alias_value":"4XGPFRNHLSQROELL","created_at":"2026-07-05T11:25:21.908507+00:00"},{"alias_kind":"pith_short_8","alias_value":"4XGPFRNH","created_at":"2026-07-05T11:25:21.908507+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10781","citing_title":"Recovering the Zipfian Distribution in Unsupervised Term Discovery","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL","json":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL.json","graph_json":"https://pith.science/api/pith-number/4XGPFRNHLSQROELL2LKQMK3QFL/graph.json","events_json":"https://pith.science/api/pith-number/4XGPFRNHLSQROELL2LKQMK3QFL/events.json","paper":"https://pith.science/paper/4XGPFRNH"},"agent_actions":{"view_html":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL","download_json":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL.json","view_paper":"https://pith.science/paper/4XGPFRNH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.17690&json=true","fetch_graph":"https://pith.science/api/pith-number/4XGPFRNHLSQROELL2LKQMK3QFL/graph.json","fetch_events":"https://pith.science/api/pith-number/4XGPFRNHLSQROELL2LKQMK3QFL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL/action/storage_attestation","attest_author":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL/action/author_attestation","sign_citation":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL/action/citation_signature","submit_replication":"https://pith.science/pith/4XGPFRNHLSQROELL2LKQMK3QFL/action/replication_record"}},"created_at":"2026-07-05T11:25:21.908507+00:00","updated_at":"2026-07-05T11:25:21.908507+00:00"}