{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:EFM5PSJSPS2IHBASZTKH7REC5X","short_pith_number":"pith:EFM5PSJS","schema_version":"1.0","canonical_sha256":"2159d7c9327cb4838412ccd47fc482edfa38bd25d280ea59b98e9579f395545d","source":{"kind":"arxiv","id":"2109.04877","version":1},"attestation_state":"computed","paper":{"title":"Efficient Test Time Adapter Ensembling for Low-resource Language Varieties","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Graham Neubig, Sebastian Ruder, Xinyi Wang, Yulia Tsvetkov","submitted_at":"2021-09-10T13:44:46Z","abstract_excerpt":"Adapters are light-weight modules that allow parameter-efficient fine-tuning of pretrained models. Specialized language and task adapters have recently been proposed to facilitate cross-lingual transfer of multilingual pretrained models (Pfeiffer et al., 2020b). However, this approach requires training a separate language adapter for every language one wishes to support, which can be impractical for languages with limited data. An intuitive solution is to use a related language adapter for the new language variety, but we observe that this solution can lead to sub-optimal performance. In this "},"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":"2109.04877","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-09-10T13:44:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c4fac4f989fd68cb2501c0a28c7157057de2d6445eed5c4539e3f3c72a9a1e6b","abstract_canon_sha256":"023b88dabd68c8ddaac1880c3567ebcffe50acfa0ea79e4b474445fca72a7998"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:13:11.070500Z","signature_b64":"VjIpqKmAQ9dYkYkNio/BsovQw2tZznpaU/II9fa8pChfSrSbC4Jzr0iRiwamMSr1g9+GrLSoQpMHCdyHASqqCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2159d7c9327cb4838412ccd47fc482edfa38bd25d280ea59b98e9579f395545d","last_reissued_at":"2026-07-05T03:13:11.070082Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:13:11.070082Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient Test Time Adapter Ensembling for Low-resource Language Varieties","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Graham Neubig, Sebastian Ruder, Xinyi Wang, Yulia Tsvetkov","submitted_at":"2021-09-10T13:44:46Z","abstract_excerpt":"Adapters are light-weight modules that allow parameter-efficient fine-tuning of pretrained models. Specialized language and task adapters have recently been proposed to facilitate cross-lingual transfer of multilingual pretrained models (Pfeiffer et al., 2020b). However, this approach requires training a separate language adapter for every language one wishes to support, which can be impractical for languages with limited data. An intuitive solution is to use a related language adapter for the new language variety, but we observe that this solution can lead to sub-optimal performance. In this "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.04877","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/2109.04877/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":"2109.04877","created_at":"2026-07-05T03:13:11.070144+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.04877v1","created_at":"2026-07-05T03:13:11.070144+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.04877","created_at":"2026-07-05T03:13:11.070144+00:00"},{"alias_kind":"pith_short_12","alias_value":"EFM5PSJSPS2I","created_at":"2026-07-05T03:13:11.070144+00:00"},{"alias_kind":"pith_short_16","alias_value":"EFM5PSJSPS2IHBAS","created_at":"2026-07-05T03:13:11.070144+00:00"},{"alias_kind":"pith_short_8","alias_value":"EFM5PSJS","created_at":"2026-07-05T03:13:11.070144+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.15523","citing_title":"An Investigation of Test-time Adaptation for Audio Classification under Background Noise","ref_index":37,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X","json":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X.json","graph_json":"https://pith.science/api/pith-number/EFM5PSJSPS2IHBASZTKH7REC5X/graph.json","events_json":"https://pith.science/api/pith-number/EFM5PSJSPS2IHBASZTKH7REC5X/events.json","paper":"https://pith.science/paper/EFM5PSJS"},"agent_actions":{"view_html":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X","download_json":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X.json","view_paper":"https://pith.science/paper/EFM5PSJS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.04877&json=true","fetch_graph":"https://pith.science/api/pith-number/EFM5PSJSPS2IHBASZTKH7REC5X/graph.json","fetch_events":"https://pith.science/api/pith-number/EFM5PSJSPS2IHBASZTKH7REC5X/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X/action/storage_attestation","attest_author":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X/action/author_attestation","sign_citation":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X/action/citation_signature","submit_replication":"https://pith.science/pith/EFM5PSJSPS2IHBASZTKH7REC5X/action/replication_record"}},"created_at":"2026-07-05T03:13:11.070144+00:00","updated_at":"2026-07-05T03:13:11.070144+00:00"}