{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:VHIKQTNHKUDFKZCOIAANLHEF7L","short_pith_number":"pith:VHIKQTNH","schema_version":"1.0","canonical_sha256":"a9d0a84da7550655644e4000d59c85fad827f58b4d5857e7edbba65f9a89848b","source":{"kind":"arxiv","id":"2107.06946","version":1},"attestation_state":"computed","paper":{"title":"Towards Quantifying the Carbon Emissions of Differentially Private Machine Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CR","authors_text":"Ahmed Zamzam, Ajinkya Mulay, Aleti Vardhan, Harshita Diddee, Krithika Ramesh, Rakshit Naidu","submitted_at":"2021-07-14T19:25:25Z","abstract_excerpt":"In recent years, machine learning techniques utilizing large-scale datasets have achieved remarkable performance. Differential privacy, by means of adding noise, provides strong privacy guarantees for such learning algorithms. The cost of differential privacy is often a reduced model accuracy and a lowered convergence speed. This paper investigates the impact of differential privacy on learning algorithms in terms of their carbon footprint due to either longer run-times or failed experiments. Through extensive experiments, further guidance is provided on choosing the noise levels which can str"},"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":"2107.06946","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2021-07-14T19:25:25Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"5fb6e234c58c516cde98432c54336b04473dc436e384ef6d2f98922a69b33c3c","abstract_canon_sha256":"2ef630b12bc0047dcd4319788631396ad8bf6096a7ce61b47065e8c75f6a2f7b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:58:08.439649Z","signature_b64":"w/miK6WpeMMZybG/BToX/KXLy0WHYddYop/IePVORY0OVgr4qka/LXUFBlHxeWVT6DJLkbBiV+pO63zCiOWWAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a9d0a84da7550655644e4000d59c85fad827f58b4d5857e7edbba65f9a89848b","last_reissued_at":"2026-07-05T02:58:08.439180Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:58:08.439180Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Quantifying the Carbon Emissions of Differentially Private Machine Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CR","authors_text":"Ahmed Zamzam, Ajinkya Mulay, Aleti Vardhan, Harshita Diddee, Krithika Ramesh, Rakshit Naidu","submitted_at":"2021-07-14T19:25:25Z","abstract_excerpt":"In recent years, machine learning techniques utilizing large-scale datasets have achieved remarkable performance. Differential privacy, by means of adding noise, provides strong privacy guarantees for such learning algorithms. The cost of differential privacy is often a reduced model accuracy and a lowered convergence speed. This paper investigates the impact of differential privacy on learning algorithms in terms of their carbon footprint due to either longer run-times or failed experiments. Through extensive experiments, further guidance is provided on choosing the noise levels which can str"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2107.06946","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/2107.06946/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":"2107.06946","created_at":"2026-07-05T02:58:08.439240+00:00"},{"alias_kind":"arxiv_version","alias_value":"2107.06946v1","created_at":"2026-07-05T02:58:08.439240+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2107.06946","created_at":"2026-07-05T02:58:08.439240+00:00"},{"alias_kind":"pith_short_12","alias_value":"VHIKQTNHKUDF","created_at":"2026-07-05T02:58:08.439240+00:00"},{"alias_kind":"pith_short_16","alias_value":"VHIKQTNHKUDFKZCO","created_at":"2026-07-05T02:58:08.439240+00:00"},{"alias_kind":"pith_short_8","alias_value":"VHIKQTNH","created_at":"2026-07-05T02:58:08.439240+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.05610","citing_title":"Towards Sustainable NLP: Insights from Benchmarking Inference Energy in Large Language Models","ref_index":29,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L","json":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L.json","graph_json":"https://pith.science/api/pith-number/VHIKQTNHKUDFKZCOIAANLHEF7L/graph.json","events_json":"https://pith.science/api/pith-number/VHIKQTNHKUDFKZCOIAANLHEF7L/events.json","paper":"https://pith.science/paper/VHIKQTNH"},"agent_actions":{"view_html":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L","download_json":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L.json","view_paper":"https://pith.science/paper/VHIKQTNH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2107.06946&json=true","fetch_graph":"https://pith.science/api/pith-number/VHIKQTNHKUDFKZCOIAANLHEF7L/graph.json","fetch_events":"https://pith.science/api/pith-number/VHIKQTNHKUDFKZCOIAANLHEF7L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L/action/storage_attestation","attest_author":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L/action/author_attestation","sign_citation":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L/action/citation_signature","submit_replication":"https://pith.science/pith/VHIKQTNHKUDFKZCOIAANLHEF7L/action/replication_record"}},"created_at":"2026-07-05T02:58:08.439240+00:00","updated_at":"2026-07-05T02:58:08.439240+00:00"}