{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:R3FCDIIURJZSJ43BUPJK2WRPTZ","short_pith_number":"pith:R3FCDIIU","schema_version":"1.0","canonical_sha256":"8eca21a1148a7324f361a3d2ad5a2f9e572e600cc24f501c90b452ea09e8678d","source":{"kind":"arxiv","id":"2412.14711","version":2},"attestation_state":"computed","paper":{"title":"ReMoE: Fully Differentiable Mixture-of-Experts with ReLU Routing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jianfei Chen, Jun Zhu, Ziteng Wang","submitted_at":"2024-12-19T10:21:20Z","abstract_excerpt":"Sparsely activated Mixture-of-Experts (MoE) models are widely adopted to scale up model capacity without increasing the computation budget. However, vanilla TopK routers are trained in a discontinuous, non-differentiable way, limiting their performance and scalability. To address this issue, we propose ReMoE, a fully differentiable MoE architecture that offers a simple yet effective drop-in replacement for the conventional TopK+Softmax routing, utilizing ReLU as the router instead. We further propose methods to regulate the router's sparsity while balancing the load among experts. ReMoE's cont"},"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":"2412.14711","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-12-19T10:21:20Z","cross_cats_sorted":[],"title_canon_sha256":"e56ef04c268b02d3941283b14c687280d99f78b920afca19fb97e95330a25095","abstract_canon_sha256":"05558398e474381c498883fa72923fa2d4dbc2a92b2885eaf76094a94d52f7c8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:21:06.153881Z","signature_b64":"AfYtiFY+S+6oi3soZfu8nGSJMT4GIS88U/JWWuJITgvm9PHcv4hStdZWtnUrSbJOOMMSxZk1T23jZbR2J+BjDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8eca21a1148a7324f361a3d2ad5a2f9e572e600cc24f501c90b452ea09e8678d","last_reissued_at":"2026-07-05T10:21:06.153408Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:21:06.153408Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ReMoE: Fully Differentiable Mixture-of-Experts with ReLU Routing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jianfei Chen, Jun Zhu, Ziteng Wang","submitted_at":"2024-12-19T10:21:20Z","abstract_excerpt":"Sparsely activated Mixture-of-Experts (MoE) models are widely adopted to scale up model capacity without increasing the computation budget. However, vanilla TopK routers are trained in a discontinuous, non-differentiable way, limiting their performance and scalability. To address this issue, we propose ReMoE, a fully differentiable MoE architecture that offers a simple yet effective drop-in replacement for the conventional TopK+Softmax routing, utilizing ReLU as the router instead. We further propose methods to regulate the router's sparsity while balancing the load among experts. ReMoE's cont"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.14711","kind":"arxiv","version":2},"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/2412.14711/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":"2412.14711","created_at":"2026-07-05T10:21:06.153478+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.14711v2","created_at":"2026-07-05T10:21:06.153478+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.14711","created_at":"2026-07-05T10:21:06.153478+00:00"},{"alias_kind":"pith_short_12","alias_value":"R3FCDIIURJZS","created_at":"2026-07-05T10:21:06.153478+00:00"},{"alias_kind":"pith_short_16","alias_value":"R3FCDIIURJZSJ43B","created_at":"2026-07-05T10:21:06.153478+00:00"},{"alias_kind":"pith_short_8","alias_value":"R3FCDIIU","created_at":"2026-07-05T10:21:06.153478+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"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":53,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26938","citing_title":"Focusing on What Matters: Saliency-Harnessing Accurate Routing for Diffusion MoE","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20544","citing_title":"Toward Calibrated Mixture-of-Experts Under Distribution Shift","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08814","citing_title":"STAR: Rethinking MoE Routing as Structure-Aware Subspace Learning","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15484","citing_title":"When Does Sparse MoE Help in Vision? The Role of Backbone Compute Leverage in Sparse Routing","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07260","citing_title":"When Are Experts Misrouted? Counterfactual Routing Analysis in Mixture-of-Experts Language Models","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ","json":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ.json","graph_json":"https://pith.science/api/pith-number/R3FCDIIURJZSJ43BUPJK2WRPTZ/graph.json","events_json":"https://pith.science/api/pith-number/R3FCDIIURJZSJ43BUPJK2WRPTZ/events.json","paper":"https://pith.science/paper/R3FCDIIU"},"agent_actions":{"view_html":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ","download_json":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ.json","view_paper":"https://pith.science/paper/R3FCDIIU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.14711&json=true","fetch_graph":"https://pith.science/api/pith-number/R3FCDIIURJZSJ43BUPJK2WRPTZ/graph.json","fetch_events":"https://pith.science/api/pith-number/R3FCDIIURJZSJ43BUPJK2WRPTZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ/action/storage_attestation","attest_author":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ/action/author_attestation","sign_citation":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ/action/citation_signature","submit_replication":"https://pith.science/pith/R3FCDIIURJZSJ43BUPJK2WRPTZ/action/replication_record"}},"created_at":"2026-07-05T10:21:06.153478+00:00","updated_at":"2026-07-05T10:21:06.153478+00:00"}