{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:CEX6QJI6HWRYWLQ2QYKBRHAYMV","short_pith_number":"pith:CEX6QJI6","schema_version":"1.0","canonical_sha256":"112fe8251e3da38b2e1a8614189c18656b16b3a86d9917ec90ada89a7a41a46d","source":{"kind":"arxiv","id":"2502.07634","version":1},"attestation_state":"computed","paper":{"title":"Efficient Distributed Training through Gradient Compression with Sparsification and Quantization Techniques","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.MM"],"primary_cat":"cs.LG","authors_text":"Shantanu Kumar, Shruti Singh","submitted_at":"2024-12-07T22:55:55Z","abstract_excerpt":"This study investigates the impact of gradient compression on distributed training performance, focusing on sparsification and quantization techniques, including top-k, DGC, and QSGD. In baseline experiments, random-k compression results in severe performance degradation, highlighting its inefficacy. In contrast, using top-k and DGC at 50 times compression yields performance improvements, reducing perplexity by up to 0.06 compared to baseline. Experiments across 1, 2, and 4 workers demonstrate that conservative sparsification can have a regularizing effect, especially for smaller models, while"},"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":"2502.07634","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-12-07T22:55:55Z","cross_cats_sorted":["cs.MM"],"title_canon_sha256":"63d30599cc04aacf7c9c67814bbf4e351c1e53fe57f6378f3b15a5a41da6766e","abstract_canon_sha256":"3e55684cc325b9606eb2311a03bc7ab04ed85742fe81d4bc197332ca6635b8f7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:12:40.266385Z","signature_b64":"CUMux4XmNJrLrG8dWH0NUCmKqXY7SM9wJSU1jc3d+up5e4K2Kms4N2rOc63ie8q854ZI9OX1dH+qav+/8zROBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"112fe8251e3da38b2e1a8614189c18656b16b3a86d9917ec90ada89a7a41a46d","last_reissued_at":"2026-07-05T10:12:40.265884Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:12:40.265884Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient Distributed Training through Gradient Compression with Sparsification and Quantization Techniques","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.MM"],"primary_cat":"cs.LG","authors_text":"Shantanu Kumar, Shruti Singh","submitted_at":"2024-12-07T22:55:55Z","abstract_excerpt":"This study investigates the impact of gradient compression on distributed training performance, focusing on sparsification and quantization techniques, including top-k, DGC, and QSGD. In baseline experiments, random-k compression results in severe performance degradation, highlighting its inefficacy. In contrast, using top-k and DGC at 50 times compression yields performance improvements, reducing perplexity by up to 0.06 compared to baseline. Experiments across 1, 2, and 4 workers demonstrate that conservative sparsification can have a regularizing effect, especially for smaller models, while"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.07634","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/2502.07634/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":"2502.07634","created_at":"2026-07-05T10:12:40.265942+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.07634v1","created_at":"2026-07-05T10:12:40.265942+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.07634","created_at":"2026-07-05T10:12:40.265942+00:00"},{"alias_kind":"pith_short_12","alias_value":"CEX6QJI6HWRY","created_at":"2026-07-05T10:12:40.265942+00:00"},{"alias_kind":"pith_short_16","alias_value":"CEX6QJI6HWRYWLQ2","created_at":"2026-07-05T10:12:40.265942+00:00"},{"alias_kind":"pith_short_8","alias_value":"CEX6QJI6","created_at":"2026-07-05T10:12:40.265942+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV","json":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV.json","graph_json":"https://pith.science/api/pith-number/CEX6QJI6HWRYWLQ2QYKBRHAYMV/graph.json","events_json":"https://pith.science/api/pith-number/CEX6QJI6HWRYWLQ2QYKBRHAYMV/events.json","paper":"https://pith.science/paper/CEX6QJI6"},"agent_actions":{"view_html":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV","download_json":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV.json","view_paper":"https://pith.science/paper/CEX6QJI6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.07634&json=true","fetch_graph":"https://pith.science/api/pith-number/CEX6QJI6HWRYWLQ2QYKBRHAYMV/graph.json","fetch_events":"https://pith.science/api/pith-number/CEX6QJI6HWRYWLQ2QYKBRHAYMV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV/action/storage_attestation","attest_author":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV/action/author_attestation","sign_citation":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV/action/citation_signature","submit_replication":"https://pith.science/pith/CEX6QJI6HWRYWLQ2QYKBRHAYMV/action/replication_record"}},"created_at":"2026-07-05T10:12:40.265942+00:00","updated_at":"2026-07-05T10:12:40.265942+00:00"}