{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:W26O6OW4HB52QH3BESCCVJJPSU","short_pith_number":"pith:W26O6OW4","schema_version":"1.0","canonical_sha256":"b6bcef3adc387ba81f6124842aa52f952896ce2aa9f2c07e2a6eaaa5fb2fcc11","source":{"kind":"arxiv","id":"2204.05149","version":1},"attestation_state":"computed","paper":{"title":"The Carbon Footprint of Machine Learning Training Will Plateau, Then Shrink","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.GL"],"primary_cat":"cs.LG","authors_text":"Chen Liang, Daniel Rothchild, David Patterson, David So, Jeff Dean, Joseph Gonzalez, Lluis-Miquel Munguia, Maud Texier, Quoc Le, Urs H\\\"olzle","submitted_at":"2022-04-11T14:30:27Z","abstract_excerpt":"Machine Learning (ML) workloads have rapidly grown in importance, but raised concerns about their carbon footprint. Four best practices can reduce ML training energy by up to 100x and CO2 emissions up to 1000x. By following best practices, overall ML energy use (across research, development, and production) held steady at <15% of Google's total energy use for the past three years. If the whole ML field were to adopt best practices, total carbon emissions from training would reduce. Hence, we recommend that ML papers include emissions explicitly to foster competition on more than just model qua"},"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":"2204.05149","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-04-11T14:30:27Z","cross_cats_sorted":["cs.AI","cs.GL"],"title_canon_sha256":"b44ef79aa15987599689a87caf6b4e63e8c0e8d255f0417dd18ddb2133dfdb00","abstract_canon_sha256":"07c141a0207e5728884d58049a3104ada08b5ed278a5cb8b399eff5c3bf0c997"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:13:17.890874Z","signature_b64":"OQdVmmeuNC7B3/EMg3Bfd8e7W4bFDBuc+Cyz0fB3Hw5zC0vLdacd1d0ioMYxHoh8H+Ehp0NWkZeZNj8YjcBsBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b6bcef3adc387ba81f6124842aa52f952896ce2aa9f2c07e2a6eaaa5fb2fcc11","last_reissued_at":"2026-07-05T04:13:17.890402Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:13:17.890402Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Carbon Footprint of Machine Learning Training Will Plateau, Then Shrink","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.GL"],"primary_cat":"cs.LG","authors_text":"Chen Liang, Daniel Rothchild, David Patterson, David So, Jeff Dean, Joseph Gonzalez, Lluis-Miquel Munguia, Maud Texier, Quoc Le, Urs H\\\"olzle","submitted_at":"2022-04-11T14:30:27Z","abstract_excerpt":"Machine Learning (ML) workloads have rapidly grown in importance, but raised concerns about their carbon footprint. Four best practices can reduce ML training energy by up to 100x and CO2 emissions up to 1000x. By following best practices, overall ML energy use (across research, development, and production) held steady at <15% of Google's total energy use for the past three years. If the whole ML field were to adopt best practices, total carbon emissions from training would reduce. Hence, we recommend that ML papers include emissions explicitly to foster competition on more than just model qua"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.05149","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/2204.05149/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":"2204.05149","created_at":"2026-07-05T04:13:17.890472+00:00"},{"alias_kind":"arxiv_version","alias_value":"2204.05149v1","created_at":"2026-07-05T04:13:17.890472+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.05149","created_at":"2026-07-05T04:13:17.890472+00:00"},{"alias_kind":"pith_short_12","alias_value":"W26O6OW4HB52","created_at":"2026-07-05T04:13:17.890472+00:00"},{"alias_kind":"pith_short_16","alias_value":"W26O6OW4HB52QH3B","created_at":"2026-07-05T04:13:17.890472+00:00"},{"alias_kind":"pith_short_8","alias_value":"W26O6OW4","created_at":"2026-07-05T04:13:17.890472+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07632","citing_title":"Evaluation of ML Resource Utilization Requires Model Life Cycle Assessment","ref_index":94,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18113","citing_title":"iPOE: Interpretable Prompt Optimization via Explanations","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27841","citing_title":"WattLayer: Get Layers Right to Estimate Inference Energy of Neural Networks","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2511.14998","citing_title":"FinCriticalED: A Visual Benchmark for Financial Fact-Level OCR","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2602.22220","citing_title":"What Makes an Ideal Quote? Recommending \"Unexpected yet Rational\" Quotations via Novelty","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13872","citing_title":"S-AI-Recursive: A Bio-Inspired and Temporal Sparse AI Architecture for Iterative, Introspective, and Energy-Frugal Reasoning","ref_index":36,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU","json":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU.json","graph_json":"https://pith.science/api/pith-number/W26O6OW4HB52QH3BESCCVJJPSU/graph.json","events_json":"https://pith.science/api/pith-number/W26O6OW4HB52QH3BESCCVJJPSU/events.json","paper":"https://pith.science/paper/W26O6OW4"},"agent_actions":{"view_html":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU","download_json":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU.json","view_paper":"https://pith.science/paper/W26O6OW4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2204.05149&json=true","fetch_graph":"https://pith.science/api/pith-number/W26O6OW4HB52QH3BESCCVJJPSU/graph.json","fetch_events":"https://pith.science/api/pith-number/W26O6OW4HB52QH3BESCCVJJPSU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU/action/storage_attestation","attest_author":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU/action/author_attestation","sign_citation":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU/action/citation_signature","submit_replication":"https://pith.science/pith/W26O6OW4HB52QH3BESCCVJJPSU/action/replication_record"}},"created_at":"2026-07-05T04:13:17.890472+00:00","updated_at":"2026-07-05T04:13:17.890472+00:00"}