{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:TL74742EQJSVLWHVCPTHIG5UNN","short_pith_number":"pith:TL74742E","schema_version":"1.0","canonical_sha256":"9affcff344826555d8f513e6741bb46b775a5e16024868831a21a954fbdf18ed","source":{"kind":"arxiv","id":"2405.14297","version":4},"attestation_state":"computed","paper":{"title":"Dynamic Mixture of Experts: An Auto-Tuning Approach for Efficient Transformer Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Tao Lin, Xiaoying Tang, Yongxin Guo, Zhaopeng Tu, Zhenglin Cheng","submitted_at":"2024-05-23T08:18:30Z","abstract_excerpt":"The Sparse Mixture of Experts (SMoE) has been widely employed to enhance the efficiency of training and inference for Transformer-based foundational models, yielding promising results.However, the performance of SMoE heavily depends on the choice of hyper-parameters, such as the number of experts and the number of experts to be activated (referred to as top-k), resulting in significant computational overhead due to the extensive model training by searching over various hyper-parameter configurations. As a remedy, we introduce the Dynamic Mixture of Experts (DynMoE) technique. DynMoE incorporat"},"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":"2405.14297","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-23T08:18:30Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f96c852ed78bdab83e740c6791be8f797e7cbfb8953e6b3c45b26419e648e9be","abstract_canon_sha256":"1eaec1fc3e699702c70785e807ffec57aec6346492581f051d9353d3dd7adad8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:10.194878Z","signature_b64":"z5dZO9MSJas6l86U1ZzDt8lBassL4mTfFt8ChAexGiqAKNXO/9GyCQDpqFHfd9G6xAM2zsIdYiMv68PBiLx/DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9affcff344826555d8f513e6741bb46b775a5e16024868831a21a954fbdf18ed","last_reissued_at":"2026-07-05T10:27:10.194261Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:10.194261Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dynamic Mixture of Experts: An Auto-Tuning Approach for Efficient Transformer Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Tao Lin, Xiaoying Tang, Yongxin Guo, Zhaopeng Tu, Zhenglin Cheng","submitted_at":"2024-05-23T08:18:30Z","abstract_excerpt":"The Sparse Mixture of Experts (SMoE) has been widely employed to enhance the efficiency of training and inference for Transformer-based foundational models, yielding promising results.However, the performance of SMoE heavily depends on the choice of hyper-parameters, such as the number of experts and the number of experts to be activated (referred to as top-k), resulting in significant computational overhead due to the extensive model training by searching over various hyper-parameter configurations. As a remedy, we introduce the Dynamic Mixture of Experts (DynMoE) technique. DynMoE incorporat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.14297","kind":"arxiv","version":4},"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/2405.14297/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":"2405.14297","created_at":"2026-07-05T10:27:10.194324+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.14297v4","created_at":"2026-07-05T10:27:10.194324+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.14297","created_at":"2026-07-05T10:27:10.194324+00:00"},{"alias_kind":"pith_short_12","alias_value":"TL74742EQJSV","created_at":"2026-07-05T10:27:10.194324+00:00"},{"alias_kind":"pith_short_16","alias_value":"TL74742EQJSVLWHV","created_at":"2026-07-05T10:27:10.194324+00:00"},{"alias_kind":"pith_short_8","alias_value":"TL74742E","created_at":"2026-07-05T10:27:10.194324+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"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":24,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08814","citing_title":"STAR: Rethinking MoE Routing as Structure-Aware Subspace Learning","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18643","citing_title":"Post-Trained MoE Can Skip Half Experts via Self-Distillation","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29982","citing_title":"Beyond Uniform Experts: Cost-Aware Expert Execution for Efficient Multi-Device MoE Inference","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.31463","citing_title":"PithTrain: A Compact and Agent-Native MoE Training System","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2502.15315","citing_title":"Tight Clusters Make Specialized Experts","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18643","citing_title":"Post-Trained MoE Can Skip Half Experts via Self-Distillation","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2509.23638","citing_title":"LayerScope: Predictive Cross-Layer Scheduling for Efficient Multi-Batch MoE Inference on Legacy Servers","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14438","citing_title":"BEAM: Binary Expert Activation Masking for Dynamic Routing in MoE","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN","json":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN.json","graph_json":"https://pith.science/api/pith-number/TL74742EQJSVLWHVCPTHIG5UNN/graph.json","events_json":"https://pith.science/api/pith-number/TL74742EQJSVLWHVCPTHIG5UNN/events.json","paper":"https://pith.science/paper/TL74742E"},"agent_actions":{"view_html":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN","download_json":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN.json","view_paper":"https://pith.science/paper/TL74742E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.14297&json=true","fetch_graph":"https://pith.science/api/pith-number/TL74742EQJSVLWHVCPTHIG5UNN/graph.json","fetch_events":"https://pith.science/api/pith-number/TL74742EQJSVLWHVCPTHIG5UNN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN/action/storage_attestation","attest_author":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN/action/author_attestation","sign_citation":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN/action/citation_signature","submit_replication":"https://pith.science/pith/TL74742EQJSVLWHVCPTHIG5UNN/action/replication_record"}},"created_at":"2026-07-05T10:27:10.194324+00:00","updated_at":"2026-07-05T10:27:10.194324+00:00"}