{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:262XDX47JSMMOJ4JFXHQ5EIDCO","short_pith_number":"pith:262XDX47","schema_version":"1.0","canonical_sha256":"d7b571df9f4c98c727892dcf0e91031391bef1cee6836e5fbf6f2c1721596838","source":{"kind":"arxiv","id":"2504.05897","version":1},"attestation_state":"computed","paper":{"title":"HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Ling Liang, Meng Li, Ru Huang, Runsheng Wang, Shuzhang Zhong, Yanfan Sun","submitted_at":"2025-04-08T10:47:37Z","abstract_excerpt":"The Mixture of Experts (MoE) architecture has demonstrated significant advantages as it enables to increase the model capacity without a proportional increase in computation. However, the large MoE model size still introduces substantial memory demands, which usually requires expert offloading on resource-constrained platforms and incurs significant overhead. Hybrid CPU-GPU inference has been proposed to leverage CPU computation to reduce expert loading overhead but faces major challenges: on one hand, the expert activation patterns of MoE models are highly unstable, rendering the fixed mappin"},"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":"2504.05897","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-04-08T10:47:37Z","cross_cats_sorted":["cs.DC"],"title_canon_sha256":"6a0d58f85fb6165142703cccb211435381cd92a820d02f6def57f76ff958a0b7","abstract_canon_sha256":"fadc8b10f4671d845386c0c790f2e27378e754b9af2132625fe6d3dcce013496"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:46:11.738731Z","signature_b64":"LC2dPBMAh12n7ypmQVBMPW0ZjMa+VFulU82RWJGQwlMRp6YuBwC6k6r3mWNnawx7zeBXhiJbsBim9u1hoB6ECw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d7b571df9f4c98c727892dcf0e91031391bef1cee6836e5fbf6f2c1721596838","last_reissued_at":"2026-07-05T10:46:11.738134Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:46:11.738134Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Ling Liang, Meng Li, Ru Huang, Runsheng Wang, Shuzhang Zhong, Yanfan Sun","submitted_at":"2025-04-08T10:47:37Z","abstract_excerpt":"The Mixture of Experts (MoE) architecture has demonstrated significant advantages as it enables to increase the model capacity without a proportional increase in computation. However, the large MoE model size still introduces substantial memory demands, which usually requires expert offloading on resource-constrained platforms and incurs significant overhead. Hybrid CPU-GPU inference has been proposed to leverage CPU computation to reduce expert loading overhead but faces major challenges: on one hand, the expert activation patterns of MoE models are highly unstable, rendering the fixed mappin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.05897","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/2504.05897/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":"2504.05897","created_at":"2026-07-05T10:46:11.738205+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.05897v1","created_at":"2026-07-05T10:46:11.738205+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.05897","created_at":"2026-07-05T10:46:11.738205+00:00"},{"alias_kind":"pith_short_12","alias_value":"262XDX47JSMM","created_at":"2026-07-05T10:46:11.738205+00:00"},{"alias_kind":"pith_short_16","alias_value":"262XDX47JSMMOJ4J","created_at":"2026-07-05T10:46:11.738205+00:00"},{"alias_kind":"pith_short_8","alias_value":"262XDX47","created_at":"2026-07-05T10:46:11.738205+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.17889","citing_title":"CoX-MoE: Coalesced Expert Execution for High-Throughput MoE Inference with AMX-Enabled CPU-GPU Co-Execution","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2508.18983","citing_title":"SMoE: An Algorithm-System Co-Design for Pushing MoE to the Edge via Expert Substitution","ref_index":19,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO","json":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO.json","graph_json":"https://pith.science/api/pith-number/262XDX47JSMMOJ4JFXHQ5EIDCO/graph.json","events_json":"https://pith.science/api/pith-number/262XDX47JSMMOJ4JFXHQ5EIDCO/events.json","paper":"https://pith.science/paper/262XDX47"},"agent_actions":{"view_html":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO","download_json":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO.json","view_paper":"https://pith.science/paper/262XDX47","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.05897&json=true","fetch_graph":"https://pith.science/api/pith-number/262XDX47JSMMOJ4JFXHQ5EIDCO/graph.json","fetch_events":"https://pith.science/api/pith-number/262XDX47JSMMOJ4JFXHQ5EIDCO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO/action/storage_attestation","attest_author":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO/action/author_attestation","sign_citation":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO/action/citation_signature","submit_replication":"https://pith.science/pith/262XDX47JSMMOJ4JFXHQ5EIDCO/action/replication_record"}},"created_at":"2026-07-05T10:46:11.738205+00:00","updated_at":"2026-07-05T10:46:11.738205+00:00"}