{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ZIQRAQSS2E5TQLH7XDSEC4HFMR","short_pith_number":"pith:ZIQRAQSS","schema_version":"1.0","canonical_sha256":"ca21104252d13b382cffb8e44170e56443ce39b7e2ad24900003dbedbc7dd9ad","source":{"kind":"arxiv","id":"2503.05447","version":2},"attestation_state":"computed","paper":{"title":"Linear-MoE: Linear Sequence Modeling Meets Mixture-of-Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.DC"],"primary_cat":"cs.LG","authors_text":"Disen Lan, Tong Zhu, Weigao Sun, Xiaoye Qu, Yu Cheng","submitted_at":"2025-03-07T14:17:45Z","abstract_excerpt":"Linear Sequence Modeling (LSM) like linear attention, state space models and linear RNNs, and Mixture-of-Experts (MoE) have recently emerged as significant architectural improvements. In this paper, we introduce Linear-MoE, a production-level system for modeling and training large-scale models that integrate LSM with MoE. Linear-MoE leverages the advantages of both LSM modules for linear-complexity sequence modeling and MoE layers for sparsely activation, aiming to offer high performance with efficient training. The Linear-MoE system comprises: 1) Modeling subsystem, which provides a unified f"},"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":"2503.05447","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-07T14:17:45Z","cross_cats_sorted":["cs.AI","cs.CL","cs.DC"],"title_canon_sha256":"da38231ed9daf863c3569dbb4945de6fc4781680c6064c9074fe01a58a1c3941","abstract_canon_sha256":"293b9bfb112a055bed67db97caa7f060c249ae08250b0b47703e57402b53bdc7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:49:14.765896Z","signature_b64":"Fc/xTRKBbUJDGkItaJDH4bwPVCq0urGIHfpvn95UEuMbIcf6qzWnS3AmpfRaAWZk5lTL2E3teEic5lkRngO6AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ca21104252d13b382cffb8e44170e56443ce39b7e2ad24900003dbedbc7dd9ad","last_reissued_at":"2026-07-05T10:49:14.765428Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:49:14.765428Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Linear-MoE: Linear Sequence Modeling Meets Mixture-of-Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.DC"],"primary_cat":"cs.LG","authors_text":"Disen Lan, Tong Zhu, Weigao Sun, Xiaoye Qu, Yu Cheng","submitted_at":"2025-03-07T14:17:45Z","abstract_excerpt":"Linear Sequence Modeling (LSM) like linear attention, state space models and linear RNNs, and Mixture-of-Experts (MoE) have recently emerged as significant architectural improvements. In this paper, we introduce Linear-MoE, a production-level system for modeling and training large-scale models that integrate LSM with MoE. Linear-MoE leverages the advantages of both LSM modules for linear-complexity sequence modeling and MoE layers for sparsely activation, aiming to offer high performance with efficient training. The Linear-MoE system comprises: 1) Modeling subsystem, which provides a unified f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.05447","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/2503.05447/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":"2503.05447","created_at":"2026-07-05T10:49:14.765489+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.05447v2","created_at":"2026-07-05T10:49:14.765489+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.05447","created_at":"2026-07-05T10:49:14.765489+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZIQRAQSS2E5T","created_at":"2026-07-05T10:49:14.765489+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZIQRAQSS2E5TQLH7","created_at":"2026-07-05T10:49:14.765489+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZIQRAQSS","created_at":"2026-07-05T10:49:14.765489+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.30562","citing_title":"Morphing into Hybrid Attention Models","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2506.13585","citing_title":"MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention","ref_index":38,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR","json":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR.json","graph_json":"https://pith.science/api/pith-number/ZIQRAQSS2E5TQLH7XDSEC4HFMR/graph.json","events_json":"https://pith.science/api/pith-number/ZIQRAQSS2E5TQLH7XDSEC4HFMR/events.json","paper":"https://pith.science/paper/ZIQRAQSS"},"agent_actions":{"view_html":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR","download_json":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR.json","view_paper":"https://pith.science/paper/ZIQRAQSS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.05447&json=true","fetch_graph":"https://pith.science/api/pith-number/ZIQRAQSS2E5TQLH7XDSEC4HFMR/graph.json","fetch_events":"https://pith.science/api/pith-number/ZIQRAQSS2E5TQLH7XDSEC4HFMR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR/action/storage_attestation","attest_author":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR/action/author_attestation","sign_citation":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR/action/citation_signature","submit_replication":"https://pith.science/pith/ZIQRAQSS2E5TQLH7XDSEC4HFMR/action/replication_record"}},"created_at":"2026-07-05T10:49:14.765489+00:00","updated_at":"2026-07-05T10:49:14.765489+00:00"}