{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:B3XWF252QIQZXRMO7FMJ2G3HUX","short_pith_number":"pith:B3XWF252","schema_version":"1.0","canonical_sha256":"0eef62ebba82219bc58ef9589d1b67a5d22727e0e64806777effcdc120da9a50","source":{"kind":"arxiv","id":"2008.08445","version":1},"attestation_state":"computed","paper":{"title":"Domain-specific Communication Optimization for Distributed DNN Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Gaoxiong Zeng, Han Tian, Hao Wang, Jiacheng Xia, Jingrong Chen, Junchen Jiang, Kai Chen, Wei Bai, Weiyan Wang, Xinchen Wan","submitted_at":"2020-08-16T09:53:21Z","abstract_excerpt":"Communication overhead poses an important obstacle to distributed DNN training and draws increasing attention in recent years. Despite continuous efforts, prior solutions such as gradient compression/reduction, compute/communication overlapping and layer-wise flow scheduling, etc., are still coarse-grained and insufficient for an efficient distributed training especially when the network is under pressure. We present DLCP, a novel solution exploiting the domain-specific properties of deep learning to optimize communication overhead of DNN training in a fine-grained manner. At its heart, DLCP c"},"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":"2008.08445","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2020-08-16T09:53:21Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e72e47a559f55860e7631d6051911ba5514f9e4ce1ee6e2f54517538d4228338","abstract_canon_sha256":"68af8bcb476def4ab98bbf40bc8af48c898532c896b494a46898fd1bd8dd3def"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:28:25.841602Z","signature_b64":"Ze7oxT0WnKJ+QtXItvGZTNd1xS9TD2Q9SE+6CdBRS4w860jkTxlzBgWfhg5Zag/yRj4/+bmAxAS6MUWmI6ngAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0eef62ebba82219bc58ef9589d1b67a5d22727e0e64806777effcdc120da9a50","last_reissued_at":"2026-07-05T01:28:25.841203Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:28:25.841203Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Domain-specific Communication Optimization for Distributed DNN Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.DC","authors_text":"Gaoxiong Zeng, Han Tian, Hao Wang, Jiacheng Xia, Jingrong Chen, Junchen Jiang, Kai Chen, Wei Bai, Weiyan Wang, Xinchen Wan","submitted_at":"2020-08-16T09:53:21Z","abstract_excerpt":"Communication overhead poses an important obstacle to distributed DNN training and draws increasing attention in recent years. Despite continuous efforts, prior solutions such as gradient compression/reduction, compute/communication overlapping and layer-wise flow scheduling, etc., are still coarse-grained and insufficient for an efficient distributed training especially when the network is under pressure. We present DLCP, a novel solution exploiting the domain-specific properties of deep learning to optimize communication overhead of DNN training in a fine-grained manner. At its heart, DLCP c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2008.08445","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/2008.08445/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":"2008.08445","created_at":"2026-07-05T01:28:25.841267+00:00"},{"alias_kind":"arxiv_version","alias_value":"2008.08445v1","created_at":"2026-07-05T01:28:25.841267+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2008.08445","created_at":"2026-07-05T01:28:25.841267+00:00"},{"alias_kind":"pith_short_12","alias_value":"B3XWF252QIQZ","created_at":"2026-07-05T01:28:25.841267+00:00"},{"alias_kind":"pith_short_16","alias_value":"B3XWF252QIQZXRMO","created_at":"2026-07-05T01:28:25.841267+00:00"},{"alias_kind":"pith_short_8","alias_value":"B3XWF252","created_at":"2026-07-05T01:28:25.841267+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.05876","citing_title":"Shesha: Opportunistic In-network Acceleration of Asynchronous Distributed Reinforcement Learning","ref_index":53,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX","json":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX.json","graph_json":"https://pith.science/api/pith-number/B3XWF252QIQZXRMO7FMJ2G3HUX/graph.json","events_json":"https://pith.science/api/pith-number/B3XWF252QIQZXRMO7FMJ2G3HUX/events.json","paper":"https://pith.science/paper/B3XWF252"},"agent_actions":{"view_html":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX","download_json":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX.json","view_paper":"https://pith.science/paper/B3XWF252","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2008.08445&json=true","fetch_graph":"https://pith.science/api/pith-number/B3XWF252QIQZXRMO7FMJ2G3HUX/graph.json","fetch_events":"https://pith.science/api/pith-number/B3XWF252QIQZXRMO7FMJ2G3HUX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX/action/storage_attestation","attest_author":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX/action/author_attestation","sign_citation":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX/action/citation_signature","submit_replication":"https://pith.science/pith/B3XWF252QIQZXRMO7FMJ2G3HUX/action/replication_record"}},"created_at":"2026-07-05T01:28:25.841267+00:00","updated_at":"2026-07-05T01:28:25.841267+00:00"}