{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:IIVPRGAMX2RBDTPANR4W3YE3EP","short_pith_number":"pith:IIVPRGAM","schema_version":"1.0","canonical_sha256":"422af8980cbea211cde06c796de09b23de59e33f352630afe8a45cd5ba9d0fff","source":{"kind":"arxiv","id":"2011.14203","version":5},"attestation_state":"computed","paper":{"title":"EdgeBERT: Sentence-Level Energy Optimizations for Latency-Aware Multi-Task NLP Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AR","authors_text":"Alexander M. Rush, Coleman Hooper, David Brooks, En-Yu Yang, Gu-Yeon Wei, Lillian Pentecost, Marco Donato, Paul N. Whatmough, Thierry Tambe, Tianyu Jia, Victor Sanh","submitted_at":"2020-11-28T19:21:47Z","abstract_excerpt":"Transformer-based language models such as BERT provide significant accuracy improvement for a multitude of natural language processing (NLP) tasks. However, their hefty computational and memory demands make them challenging to deploy to resource-constrained edge platforms with strict latency requirements. We present EdgeBERT, an in-depth algorithm-hardware co-design for latency-aware energy optimization for multi-task NLP. EdgeBERT employs entropy-based early exit predication in order to perform dynamic voltage-frequency scaling (DVFS), at a sentence granularity, for minimal energy consumption"},"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":"2011.14203","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2020-11-28T19:21:47Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"0ed23d9c15b33e59b0e17f11f9b0c7d951ffa7d8a53194aa60273587b754b155","abstract_canon_sha256":"f56f11a5336a732c388459bc16beb0adb98cff2100593eada4fdc68e3218d560"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:11:27.018191Z","signature_b64":"6Emp6UbPmt0rqtF7KBfiDakWdgNnj3WJ/ZfP1sJzpMm4S+dHUrHFbzLwiUJB4QcqXbMQhbiJZXyTjGUQIb4LAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"422af8980cbea211cde06c796de09b23de59e33f352630afe8a45cd5ba9d0fff","last_reissued_at":"2026-07-05T03:11:27.017772Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:11:27.017772Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EdgeBERT: Sentence-Level Energy Optimizations for Latency-Aware Multi-Task NLP Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AR","authors_text":"Alexander M. Rush, Coleman Hooper, David Brooks, En-Yu Yang, Gu-Yeon Wei, Lillian Pentecost, Marco Donato, Paul N. Whatmough, Thierry Tambe, Tianyu Jia, Victor Sanh","submitted_at":"2020-11-28T19:21:47Z","abstract_excerpt":"Transformer-based language models such as BERT provide significant accuracy improvement for a multitude of natural language processing (NLP) tasks. However, their hefty computational and memory demands make them challenging to deploy to resource-constrained edge platforms with strict latency requirements. We present EdgeBERT, an in-depth algorithm-hardware co-design for latency-aware energy optimization for multi-task NLP. EdgeBERT employs entropy-based early exit predication in order to perform dynamic voltage-frequency scaling (DVFS), at a sentence granularity, for minimal energy consumption"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2011.14203","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/2011.14203/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":"2011.14203","created_at":"2026-07-05T03:11:27.017829+00:00"},{"alias_kind":"arxiv_version","alias_value":"2011.14203v5","created_at":"2026-07-05T03:11:27.017829+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2011.14203","created_at":"2026-07-05T03:11:27.017829+00:00"},{"alias_kind":"pith_short_12","alias_value":"IIVPRGAMX2RB","created_at":"2026-07-05T03:11:27.017829+00:00"},{"alias_kind":"pith_short_16","alias_value":"IIVPRGAMX2RBDTPA","created_at":"2026-07-05T03:11:27.017829+00:00"},{"alias_kind":"pith_short_8","alias_value":"IIVPRGAM","created_at":"2026-07-05T03:11:27.017829+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.05527","citing_title":"EdgeServing: Deadline-Aware Multi-DNN Serving at the Edge","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP","json":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP.json","graph_json":"https://pith.science/api/pith-number/IIVPRGAMX2RBDTPANR4W3YE3EP/graph.json","events_json":"https://pith.science/api/pith-number/IIVPRGAMX2RBDTPANR4W3YE3EP/events.json","paper":"https://pith.science/paper/IIVPRGAM"},"agent_actions":{"view_html":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP","download_json":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP.json","view_paper":"https://pith.science/paper/IIVPRGAM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2011.14203&json=true","fetch_graph":"https://pith.science/api/pith-number/IIVPRGAMX2RBDTPANR4W3YE3EP/graph.json","fetch_events":"https://pith.science/api/pith-number/IIVPRGAMX2RBDTPANR4W3YE3EP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP/action/storage_attestation","attest_author":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP/action/author_attestation","sign_citation":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP/action/citation_signature","submit_replication":"https://pith.science/pith/IIVPRGAMX2RBDTPANR4W3YE3EP/action/replication_record"}},"created_at":"2026-07-05T03:11:27.017829+00:00","updated_at":"2026-07-05T03:11:27.017829+00:00"}