{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:AIU3XYORINTL7D5SAJEB3CNNTY","short_pith_number":"pith:AIU3XYOR","schema_version":"1.0","canonical_sha256":"0229bbe1d14366bf8fb202481d89ad9e07df0def90402c0c7558a611410d04c7","source":{"kind":"arxiv","id":"2303.01610","version":1},"attestation_state":"computed","paper":{"title":"Sparse MoE as the New Dropout: Scaling Dense and Self-Slimmable Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ajay Jaiswal, Shiwei Liu, Tianlong Chen, Zhangyang Wang, Zhenyu Zhang","submitted_at":"2023-03-02T22:12:51Z","abstract_excerpt":"Despite their remarkable achievement, gigantic transformers encounter significant drawbacks, including exorbitant computational and memory footprints during training, as well as severe collapse evidenced by a high degree of parameter redundancy. Sparsely-activated Mixture-of-Experts (SMoEs) have shown promise to mitigate the issue of training efficiency, yet they are prone to (1) redundant experts due to representational collapse; and (2) poor expert scalability for inference and downstream fine-tuning, primarily due to overfitting of the learned routing policy to the number of activated exper"},"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":"2303.01610","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-03-02T22:12:51Z","cross_cats_sorted":[],"title_canon_sha256":"ae4b4d21d440eab7b3384403c7e589c1b9a586768ab5fd47cd2c085906853c78","abstract_canon_sha256":"8ebc4cfb3e8602ce70d9dec15565cd2ad20e9e2cfb43ff7df6c0920f9713f1bf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:47:36.447822Z","signature_b64":"JGSQ9159kmkwCx0hCkbcIA+7rsx/kdxwcpTNdmR8lfnqnYFpgF2d6WDW4qEigSj+TEnrKVvh6Z1MWd84L59BBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0229bbe1d14366bf8fb202481d89ad9e07df0def90402c0c7558a611410d04c7","last_reissued_at":"2026-07-05T05:47:36.447352Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:47:36.447352Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sparse MoE as the New Dropout: Scaling Dense and Self-Slimmable Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ajay Jaiswal, Shiwei Liu, Tianlong Chen, Zhangyang Wang, Zhenyu Zhang","submitted_at":"2023-03-02T22:12:51Z","abstract_excerpt":"Despite their remarkable achievement, gigantic transformers encounter significant drawbacks, including exorbitant computational and memory footprints during training, as well as severe collapse evidenced by a high degree of parameter redundancy. Sparsely-activated Mixture-of-Experts (SMoEs) have shown promise to mitigate the issue of training efficiency, yet they are prone to (1) redundant experts due to representational collapse; and (2) poor expert scalability for inference and downstream fine-tuning, primarily due to overfitting of the learned routing policy to the number of activated exper"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.01610","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/2303.01610/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":"2303.01610","created_at":"2026-07-05T05:47:36.447402+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.01610v1","created_at":"2026-07-05T05:47:36.447402+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.01610","created_at":"2026-07-05T05:47:36.447402+00:00"},{"alias_kind":"pith_short_12","alias_value":"AIU3XYORINTL","created_at":"2026-07-05T05:47:36.447402+00:00"},{"alias_kind":"pith_short_16","alias_value":"AIU3XYORINTL7D5S","created_at":"2026-07-05T05:47:36.447402+00:00"},{"alias_kind":"pith_short_8","alias_value":"AIU3XYOR","created_at":"2026-07-05T05:47:36.447402+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.04378","citing_title":"DLLG: Dynamic Logit-Level Gating of LLM Experts","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2411.08982","citing_title":"Lynx: Enabling Efficient MoE Inference through Dynamic Batch-Aware Expert Selection","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2512.23070","citing_title":"FLEX-MoE: Federated Mixture-of-Experts with Load-balanced Expert Assignment for Edge Computing","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06542","citing_title":"Does a Global Perspective Help Prune Sparse MoEs Elegantly?","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY","json":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY.json","graph_json":"https://pith.science/api/pith-number/AIU3XYORINTL7D5SAJEB3CNNTY/graph.json","events_json":"https://pith.science/api/pith-number/AIU3XYORINTL7D5SAJEB3CNNTY/events.json","paper":"https://pith.science/paper/AIU3XYOR"},"agent_actions":{"view_html":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY","download_json":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY.json","view_paper":"https://pith.science/paper/AIU3XYOR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.01610&json=true","fetch_graph":"https://pith.science/api/pith-number/AIU3XYORINTL7D5SAJEB3CNNTY/graph.json","fetch_events":"https://pith.science/api/pith-number/AIU3XYORINTL7D5SAJEB3CNNTY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY/action/storage_attestation","attest_author":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY/action/author_attestation","sign_citation":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY/action/citation_signature","submit_replication":"https://pith.science/pith/AIU3XYORINTL7D5SAJEB3CNNTY/action/replication_record"}},"created_at":"2026-07-05T05:47:36.447402+00:00","updated_at":"2026-07-05T05:47:36.447402+00:00"}