{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:OFVGI5TXBPEJ5DTKIOOZN2JFUI","short_pith_number":"pith:OFVGI5TX","schema_version":"1.0","canonical_sha256":"716a6476770bc89e8e6a439d96e925a211abbda4d5e78621de46221146e5d151","source":{"kind":"arxiv","id":"2407.02610","version":2},"attestation_state":"computed","paper":{"title":"Towards Federated Learning with On-device Training and Communication in 8-bit Floating Point","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Axel Berg, Bokun Wang, Chuteng Zhou, Durmus Alp Emre Acar","submitted_at":"2024-07-02T18:55:58Z","abstract_excerpt":"Recent work has shown that 8-bit floating point (FP8) can be used for efficiently training neural networks with reduced computational cost compared to training in FP32/FP16. In this work, we investigate the use of FP8 training in a federated learning context. This approach brings not only the usual benefits of FP8 which are desirable for on-device training at the edge, but also reduces client-server communication costs due to significant weight compression. We present a novel method for combining FP8 client training while maintaining a global FP32 server model and provide convergence analysis."},"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":"2407.02610","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-02T18:55:58Z","cross_cats_sorted":["cs.DC"],"title_canon_sha256":"447f6e23c5c699ff97fd6664369501c03c11dcd3c420e13c0f6f1deeb8ba59cc","abstract_canon_sha256":"406974ca2532581828c6d5d1fb24f01f39feff609454f474a06ed9c65a4c780d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:45:28.329537Z","signature_b64":"yHUWm62eZLMDebc9Q+CtGrQTb63+Q0RNo05QaxKPMf/BAcBFe8tvJjy7tF2RU5VH0Q/yFZMk9hHaDjAXwN43Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"716a6476770bc89e8e6a439d96e925a211abbda4d5e78621de46221146e5d151","last_reissued_at":"2026-07-05T11:45:28.328698Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:45:28.328698Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Federated Learning with On-device Training and Communication in 8-bit Floating Point","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Axel Berg, Bokun Wang, Chuteng Zhou, Durmus Alp Emre Acar","submitted_at":"2024-07-02T18:55:58Z","abstract_excerpt":"Recent work has shown that 8-bit floating point (FP8) can be used for efficiently training neural networks with reduced computational cost compared to training in FP32/FP16. In this work, we investigate the use of FP8 training in a federated learning context. This approach brings not only the usual benefits of FP8 which are desirable for on-device training at the edge, but also reduces client-server communication costs due to significant weight compression. We present a novel method for combining FP8 client training while maintaining a global FP32 server model and provide convergence analysis."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.02610","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/2407.02610/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":"2407.02610","created_at":"2026-07-05T11:45:28.328779+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.02610v2","created_at":"2026-07-05T11:45:28.328779+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.02610","created_at":"2026-07-05T11:45:28.328779+00:00"},{"alias_kind":"pith_short_12","alias_value":"OFVGI5TXBPEJ","created_at":"2026-07-05T11:45:28.328779+00:00"},{"alias_kind":"pith_short_16","alias_value":"OFVGI5TXBPEJ5DTK","created_at":"2026-07-05T11:45:28.328779+00:00"},{"alias_kind":"pith_short_8","alias_value":"OFVGI5TX","created_at":"2026-07-05T11:45:28.328779+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.11982","citing_title":"FedHQ: Hybrid Runtime Quantization for Federated Learning","ref_index":13,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI","json":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI.json","graph_json":"https://pith.science/api/pith-number/OFVGI5TXBPEJ5DTKIOOZN2JFUI/graph.json","events_json":"https://pith.science/api/pith-number/OFVGI5TXBPEJ5DTKIOOZN2JFUI/events.json","paper":"https://pith.science/paper/OFVGI5TX"},"agent_actions":{"view_html":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI","download_json":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI.json","view_paper":"https://pith.science/paper/OFVGI5TX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.02610&json=true","fetch_graph":"https://pith.science/api/pith-number/OFVGI5TXBPEJ5DTKIOOZN2JFUI/graph.json","fetch_events":"https://pith.science/api/pith-number/OFVGI5TXBPEJ5DTKIOOZN2JFUI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI/action/storage_attestation","attest_author":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI/action/author_attestation","sign_citation":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI/action/citation_signature","submit_replication":"https://pith.science/pith/OFVGI5TXBPEJ5DTKIOOZN2JFUI/action/replication_record"}},"created_at":"2026-07-05T11:45:28.328779+00:00","updated_at":"2026-07-05T11:45:28.328779+00:00"}