{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:4TH77EWDUU4JT55XWE3FIF4KEI","short_pith_number":"pith:4TH77EWD","schema_version":"1.0","canonical_sha256":"e4cfff92c3a53899f7b7b13654178a2228cc8e9366b1ee991ff99061940734f3","source":{"kind":"arxiv","id":"2010.05904","version":1},"attestation_state":"computed","paper":{"title":"Multi-Stage Pre-training for Low-Resource Domain Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Anthony Ferritto, Avirup Sil, Efsun Sarioglu Kayi, Md Arafat Sultan, Radu Florian, Revanth Gangi Reddy, Rong Zhang, Salim Roukos, Todd Ward, Vittorio Castelli","submitted_at":"2020-10-12T17:57:00Z","abstract_excerpt":"Transfer learning techniques are particularly useful in NLP tasks where a sizable amount of high-quality annotated data is difficult to obtain. Current approaches directly adapt a pre-trained language model (LM) on in-domain text before fine-tuning to downstream tasks. We show that extending the vocabulary of the LM with domain-specific terms leads to further gains. To a bigger effect, we utilize structure in the unlabeled data to create auxiliary synthetic tasks, which helps the LM transfer to downstream tasks. We apply these approaches incrementally on a pre-trained Roberta-large LM and show"},"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":"2010.05904","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-10-12T17:57:00Z","cross_cats_sorted":[],"title_canon_sha256":"5d45da4876116e53b87fa634f9d7793211ce8e4269fe27ef76552260ba5959ea","abstract_canon_sha256":"64139c81c0de7fbf1989760bd5b214056c13c5df3e7b05cd68cb01cd51b8839a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:42:11.583558Z","signature_b64":"vIvtkDVVHLm3yMLGb0nnXZiaSBm4HAE/TMPdCJynLn1slkQ4Bd7Yc/e8MkME71pZ2xXF7QdOOB4aWgK4kTGJBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e4cfff92c3a53899f7b7b13654178a2228cc8e9366b1ee991ff99061940734f3","last_reissued_at":"2026-07-05T01:42:11.583117Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:42:11.583117Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multi-Stage Pre-training for Low-Resource Domain Adaptation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Anthony Ferritto, Avirup Sil, Efsun Sarioglu Kayi, Md Arafat Sultan, Radu Florian, Revanth Gangi Reddy, Rong Zhang, Salim Roukos, Todd Ward, Vittorio Castelli","submitted_at":"2020-10-12T17:57:00Z","abstract_excerpt":"Transfer learning techniques are particularly useful in NLP tasks where a sizable amount of high-quality annotated data is difficult to obtain. Current approaches directly adapt a pre-trained language model (LM) on in-domain text before fine-tuning to downstream tasks. We show that extending the vocabulary of the LM with domain-specific terms leads to further gains. To a bigger effect, we utilize structure in the unlabeled data to create auxiliary synthetic tasks, which helps the LM transfer to downstream tasks. We apply these approaches incrementally on a pre-trained Roberta-large LM and show"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.05904","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/2010.05904/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":"2010.05904","created_at":"2026-07-05T01:42:11.583167+00:00"},{"alias_kind":"arxiv_version","alias_value":"2010.05904v1","created_at":"2026-07-05T01:42:11.583167+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.05904","created_at":"2026-07-05T01:42:11.583167+00:00"},{"alias_kind":"pith_short_12","alias_value":"4TH77EWDUU4J","created_at":"2026-07-05T01:42:11.583167+00:00"},{"alias_kind":"pith_short_16","alias_value":"4TH77EWDUU4JT55X","created_at":"2026-07-05T01:42:11.583167+00:00"},{"alias_kind":"pith_short_8","alias_value":"4TH77EWD","created_at":"2026-07-05T01:42:11.583167+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI","json":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI.json","graph_json":"https://pith.science/api/pith-number/4TH77EWDUU4JT55XWE3FIF4KEI/graph.json","events_json":"https://pith.science/api/pith-number/4TH77EWDUU4JT55XWE3FIF4KEI/events.json","paper":"https://pith.science/paper/4TH77EWD"},"agent_actions":{"view_html":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI","download_json":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI.json","view_paper":"https://pith.science/paper/4TH77EWD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2010.05904&json=true","fetch_graph":"https://pith.science/api/pith-number/4TH77EWDUU4JT55XWE3FIF4KEI/graph.json","fetch_events":"https://pith.science/api/pith-number/4TH77EWDUU4JT55XWE3FIF4KEI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI/action/storage_attestation","attest_author":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI/action/author_attestation","sign_citation":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI/action/citation_signature","submit_replication":"https://pith.science/pith/4TH77EWDUU4JT55XWE3FIF4KEI/action/replication_record"}},"created_at":"2026-07-05T01:42:11.583167+00:00","updated_at":"2026-07-05T01:42:11.583167+00:00"}