{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:UPAPWNVJYU5QIPF4XJD6CGBMYV","short_pith_number":"pith:UPAPWNVJ","schema_version":"1.0","canonical_sha256":"a3c0fb36a9c53b043cbcba47e1182cc57be8066ffcf7451baef98c257f3880f5","source":{"kind":"arxiv","id":"2001.04246","version":2},"attestation_state":"computed","paper":{"title":"AdaBERT: Task-Adaptive BERT Compression with Differentiable Neural Architecture Search","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Bofang Li, Bolin Ding, Daoyuan Chen, Hongbo Deng, Jingren Zhou, Jun Huang, Minghui Qiu, Wei Lin, Yaliang Li, Zhen Wang","submitted_at":"2020-01-13T14:03:26Z","abstract_excerpt":"Large pre-trained language models such as BERT have shown their effectiveness in various natural language processing tasks. However, the huge parameter size makes them difficult to be deployed in real-time applications that require quick inference with limited resources. Existing methods compress BERT into small models while such compression is task-independent, i.e., the same compressed BERT for all different downstream tasks. Motivated by the necessity and benefits of task-oriented BERT compression, we propose a novel compression method, AdaBERT, that leverages differentiable Neural Architec"},"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":"2001.04246","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-01-13T14:03:26Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"17d4e864aaaac73207f232c2e5b9c7434069bcf15bf6e837f4e14787fbcff264","abstract_canon_sha256":"15405d7aa82f1e4da087b484681fe0c6db665bb4b3f9eca99c916a37353dfd39"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:08:47.880633Z","signature_b64":"VpMHNJwa31hyZ42Ib3AfHlWJdSx3kYMs6v3LtnI6IwvgpQLdA+IHj2x04H61puT8qs0OOM97HRnvFZ/e8f4qAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a3c0fb36a9c53b043cbcba47e1182cc57be8066ffcf7451baef98c257f3880f5","last_reissued_at":"2026-07-05T02:08:47.879980Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:08:47.879980Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AdaBERT: Task-Adaptive BERT Compression with Differentiable Neural Architecture Search","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Bofang Li, Bolin Ding, Daoyuan Chen, Hongbo Deng, Jingren Zhou, Jun Huang, Minghui Qiu, Wei Lin, Yaliang Li, Zhen Wang","submitted_at":"2020-01-13T14:03:26Z","abstract_excerpt":"Large pre-trained language models such as BERT have shown their effectiveness in various natural language processing tasks. However, the huge parameter size makes them difficult to be deployed in real-time applications that require quick inference with limited resources. Existing methods compress BERT into small models while such compression is task-independent, i.e., the same compressed BERT for all different downstream tasks. Motivated by the necessity and benefits of task-oriented BERT compression, we propose a novel compression method, AdaBERT, that leverages differentiable Neural Architec"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2001.04246","kind":"arxiv","version":2},"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/2001.04246/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":"2001.04246","created_at":"2026-07-05T02:08:47.880053+00:00"},{"alias_kind":"arxiv_version","alias_value":"2001.04246v2","created_at":"2026-07-05T02:08:47.880053+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2001.04246","created_at":"2026-07-05T02:08:47.880053+00:00"},{"alias_kind":"pith_short_12","alias_value":"UPAPWNVJYU5Q","created_at":"2026-07-05T02:08:47.880053+00:00"},{"alias_kind":"pith_short_16","alias_value":"UPAPWNVJYU5QIPF4","created_at":"2026-07-05T02:08:47.880053+00:00"},{"alias_kind":"pith_short_8","alias_value":"UPAPWNVJ","created_at":"2026-07-05T02:08:47.880053+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.08568","citing_title":"Different Prompts, Different Ranks: Prompt-aware Dynamic Rank Selection for SVD-based LLM Compression","ref_index":26,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV","json":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV.json","graph_json":"https://pith.science/api/pith-number/UPAPWNVJYU5QIPF4XJD6CGBMYV/graph.json","events_json":"https://pith.science/api/pith-number/UPAPWNVJYU5QIPF4XJD6CGBMYV/events.json","paper":"https://pith.science/paper/UPAPWNVJ"},"agent_actions":{"view_html":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV","download_json":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV.json","view_paper":"https://pith.science/paper/UPAPWNVJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2001.04246&json=true","fetch_graph":"https://pith.science/api/pith-number/UPAPWNVJYU5QIPF4XJD6CGBMYV/graph.json","fetch_events":"https://pith.science/api/pith-number/UPAPWNVJYU5QIPF4XJD6CGBMYV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV/action/storage_attestation","attest_author":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV/action/author_attestation","sign_citation":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV/action/citation_signature","submit_replication":"https://pith.science/pith/UPAPWNVJYU5QIPF4XJD6CGBMYV/action/replication_record"}},"created_at":"2026-07-05T02:08:47.880053+00:00","updated_at":"2026-07-05T02:08:47.880053+00:00"}