{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:3TUE3AJ2EKLGMYACTPERP6RY7P","short_pith_number":"pith:3TUE3AJ2","schema_version":"1.0","canonical_sha256":"dce84d813a22966660029bc917fa38fbd7a81f2df6727c9130535edc5d5cef77","source":{"kind":"arxiv","id":"2410.07348","version":1},"attestation_state":"computed","paper":{"title":"MoE++: Accelerating Mixture-of-Experts Methods with Zero-Computation Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bo Zhu, Li Yuan, Peng Jin, Shuicheng Yan","submitted_at":"2024-10-09T18:01:27Z","abstract_excerpt":"In this work, we aim to simultaneously enhance the effectiveness and efficiency of Mixture-of-Experts (MoE) methods. To achieve this, we propose MoE++, a general and heterogeneous MoE framework that integrates both Feed-Forward Network~(FFN) and zero-computation experts. Specifically, we introduce three types of zero-computation experts: the zero expert, copy expert, and constant expert, which correspond to discard, skip, and replace operations, respectively. This design offers three key advantages: (i) Low Computing Overhead: Unlike the uniform mixing mechanism for all tokens within vanilla M"},"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":"2410.07348","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-09T18:01:27Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"de0a391c1bba3b3a58b1c7d917ce8461029e5cb0a97533d23d36c74f355cf7d2","abstract_canon_sha256":"93941d78c76374e1a36948dc389f99dce4b57abf1511a8471f8f2ae3a914196a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:18:34.917995Z","signature_b64":"2iC/SxE5DBx8HPHyyi4vLUJUGr+HEpoYRCURN7BHaNoPl+NWbRZXrpWe7A8Udh1FoJeQ44+tcA3/blkz19axDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dce84d813a22966660029bc917fa38fbd7a81f2df6727c9130535edc5d5cef77","last_reissued_at":"2026-07-05T09:18:34.917566Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:18:34.917566Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MoE++: Accelerating Mixture-of-Experts Methods with Zero-Computation Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bo Zhu, Li Yuan, Peng Jin, Shuicheng Yan","submitted_at":"2024-10-09T18:01:27Z","abstract_excerpt":"In this work, we aim to simultaneously enhance the effectiveness and efficiency of Mixture-of-Experts (MoE) methods. To achieve this, we propose MoE++, a general and heterogeneous MoE framework that integrates both Feed-Forward Network~(FFN) and zero-computation experts. Specifically, we introduce three types of zero-computation experts: the zero expert, copy expert, and constant expert, which correspond to discard, skip, and replace operations, respectively. This design offers three key advantages: (i) Low Computing Overhead: Unlike the uniform mixing mechanism for all tokens within vanilla M"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.07348","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/2410.07348/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":"2410.07348","created_at":"2026-07-05T09:18:34.917625+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.07348v1","created_at":"2026-07-05T09:18:34.917625+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.07348","created_at":"2026-07-05T09:18:34.917625+00:00"},{"alias_kind":"pith_short_12","alias_value":"3TUE3AJ2EKLG","created_at":"2026-07-05T09:18:34.917625+00:00"},{"alias_kind":"pith_short_16","alias_value":"3TUE3AJ2EKLGMYAC","created_at":"2026-07-05T09:18:34.917625+00:00"},{"alias_kind":"pith_short_8","alias_value":"3TUE3AJ2","created_at":"2026-07-05T09:18:34.917625+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26287","citing_title":"GeMoE: Gating Entropy is All You Need for Uncertainty-aware Adaptive Routing in MoE-based Large Vision-Language Models","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00371","citing_title":"MEPA: Multi-Scale Representation Alignment for Visual Autoregressive Modeling with Mixture of Experts","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18643","citing_title":"Post-Trained MoE Can Skip Half Experts via Self-Distillation","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18643","citing_title":"Post-Trained MoE Can Skip Half Experts via Self-Distillation","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2602.12708","citing_title":"Mixture of Predefined Experts: Maximizing Data Usage on Vertical Federated Learning","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14438","citing_title":"BEAM: Binary Expert Activation Masking for Dynamic Routing in MoE","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08585","citing_title":"PromptDx: Differentiable Prompt Tuning for Multimodal In-Context Alzheimer's Diagnosis","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P","json":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P.json","graph_json":"https://pith.science/api/pith-number/3TUE3AJ2EKLGMYACTPERP6RY7P/graph.json","events_json":"https://pith.science/api/pith-number/3TUE3AJ2EKLGMYACTPERP6RY7P/events.json","paper":"https://pith.science/paper/3TUE3AJ2"},"agent_actions":{"view_html":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P","download_json":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P.json","view_paper":"https://pith.science/paper/3TUE3AJ2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.07348&json=true","fetch_graph":"https://pith.science/api/pith-number/3TUE3AJ2EKLGMYACTPERP6RY7P/graph.json","fetch_events":"https://pith.science/api/pith-number/3TUE3AJ2EKLGMYACTPERP6RY7P/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P/action/storage_attestation","attest_author":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P/action/author_attestation","sign_citation":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P/action/citation_signature","submit_replication":"https://pith.science/pith/3TUE3AJ2EKLGMYACTPERP6RY7P/action/replication_record"}},"created_at":"2026-07-05T09:18:34.917625+00:00","updated_at":"2026-07-05T09:18:34.917625+00:00"}