{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:RBQSXZETQISYD7WBJ7DDR72VNB","short_pith_number":"pith:RBQSXZET","schema_version":"1.0","canonical_sha256":"88612be493822581fec14fc638ff556876153165a0f35beff3270eaf8167fdcc","source":{"kind":"arxiv","id":"2310.02410","version":1},"attestation_state":"computed","paper":{"title":"Mixture of Quantized Experts (MoQE): Complementary Effect of Low-bit Quantization and Robustness","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Hany Hassan Awadalla, Raffy Fahim, Young Jin Kim","submitted_at":"2023-10-03T20:11:23Z","abstract_excerpt":"Large Mixture of Experts (MoE) models could achieve state-of-the-art quality on various language tasks, including machine translation task, thanks to the efficient model scaling capability with expert parallelism. However, it has brought a fundamental issue of larger memory consumption and increased memory bandwidth bottleneck at deployment time. In this paper, we propose Mixture of Quantized Experts (MoQE) which is a simple weight-only quantization method applying ultra low-bit down to 2-bit quantizations only to expert weights for mitigating the increased memory and latency issues of MoE mod"},"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":"2310.02410","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-03T20:11:23Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"01973c978a6daf3605f738285352d5934027032b15f82cc9d5fe2e9258f88526","abstract_canon_sha256":"5f98454e791ee391065a42ce2e358d82a3033de41a6f971494c1e454251757ea"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:56:56.922484Z","signature_b64":"UyhtiXT5u5wSwutdb59Rv2Q83Yt8EZZnNaS/sR06dX1lAr9Ft1gCFdbBjFIU6FwYMSw17jeEylGC1dmUNisEDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"88612be493822581fec14fc638ff556876153165a0f35beff3270eaf8167fdcc","last_reissued_at":"2026-07-05T06:56:56.922100Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:56:56.922100Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mixture of Quantized Experts (MoQE): Complementary Effect of Low-bit Quantization and Robustness","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Hany Hassan Awadalla, Raffy Fahim, Young Jin Kim","submitted_at":"2023-10-03T20:11:23Z","abstract_excerpt":"Large Mixture of Experts (MoE) models could achieve state-of-the-art quality on various language tasks, including machine translation task, thanks to the efficient model scaling capability with expert parallelism. However, it has brought a fundamental issue of larger memory consumption and increased memory bandwidth bottleneck at deployment time. In this paper, we propose Mixture of Quantized Experts (MoQE) which is a simple weight-only quantization method applying ultra low-bit down to 2-bit quantizations only to expert weights for mitigating the increased memory and latency issues of MoE mod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.02410","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/2310.02410/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":"2310.02410","created_at":"2026-07-05T06:56:56.922156+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.02410v1","created_at":"2026-07-05T06:56:56.922156+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.02410","created_at":"2026-07-05T06:56:56.922156+00:00"},{"alias_kind":"pith_short_12","alias_value":"RBQSXZETQISY","created_at":"2026-07-05T06:56:56.922156+00:00"},{"alias_kind":"pith_short_16","alias_value":"RBQSXZETQISYD7WB","created_at":"2026-07-05T06:56:56.922156+00:00"},{"alias_kind":"pith_short_8","alias_value":"RBQSXZET","created_at":"2026-07-05T06:56:56.922156+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.02715","citing_title":"FluxMoE: Decoupling Expert Residency for High-Performance MoE Serving","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB","json":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB.json","graph_json":"https://pith.science/api/pith-number/RBQSXZETQISYD7WBJ7DDR72VNB/graph.json","events_json":"https://pith.science/api/pith-number/RBQSXZETQISYD7WBJ7DDR72VNB/events.json","paper":"https://pith.science/paper/RBQSXZET"},"agent_actions":{"view_html":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB","download_json":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB.json","view_paper":"https://pith.science/paper/RBQSXZET","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.02410&json=true","fetch_graph":"https://pith.science/api/pith-number/RBQSXZETQISYD7WBJ7DDR72VNB/graph.json","fetch_events":"https://pith.science/api/pith-number/RBQSXZETQISYD7WBJ7DDR72VNB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB/action/storage_attestation","attest_author":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB/action/author_attestation","sign_citation":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB/action/citation_signature","submit_replication":"https://pith.science/pith/RBQSXZETQISYD7WBJ7DDR72VNB/action/replication_record"}},"created_at":"2026-07-05T06:56:56.922156+00:00","updated_at":"2026-07-05T06:56:56.922156+00:00"}