{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BZKGOXZUT6FQUF7OM7X5UOQFW5","short_pith_number":"pith:BZKGOXZU","schema_version":"1.0","canonical_sha256":"0e54675f349f8b0a17ee67efda3a05b76dabbe609e0390d21765a1d3bf2cfca9","source":{"kind":"arxiv","id":"2508.17137","version":1},"attestation_state":"computed","paper":{"title":"MoE-Beyond: Learning-Based Expert Activation Prediction on Edge Devices","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Arush Mehrotra, Nishant Gavhane, Peter Proenca, Rohit Chawla","submitted_at":"2025-08-23T20:28:32Z","abstract_excerpt":"The deployment of large-scale Mixture-of-Experts (MoE) models on edge devices presents significant challenges due to memory constraints. While MoE architectures enable efficient utilization of computational resources by activating only a subset of experts per inference, they require careful memory management to operate efficiently in resource-constrained environments. Traditional heuristic-based expert caching strategies such as MoE-Infinity struggle to maintain high cache hit rates as models parameters scale. In this work, we introduce MoE-Beyond, a learning-based expert activation predictor "},"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":"2508.17137","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-23T20:28:32Z","cross_cats_sorted":[],"title_canon_sha256":"f700d1748f252d3cb85316cb1669f674257b1230f459bdc000ccfe26cc1a8b98","abstract_canon_sha256":"284d1b34e33d6854df862149aa425f2d5c20b0db5f177c40902043782dd9205f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:58:10.106903Z","signature_b64":"W4mn2Iur1nbO0uivfuv0R5Au466ZUu9KJ+dETgIokNMan7utC12RLJ9s3sIUI1jbmXAwVlTz+FbCEYGcipyPBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0e54675f349f8b0a17ee67efda3a05b76dabbe609e0390d21765a1d3bf2cfca9","last_reissued_at":"2026-07-05T11:58:10.106484Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:58:10.106484Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MoE-Beyond: Learning-Based Expert Activation Prediction on Edge Devices","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Arush Mehrotra, Nishant Gavhane, Peter Proenca, Rohit Chawla","submitted_at":"2025-08-23T20:28:32Z","abstract_excerpt":"The deployment of large-scale Mixture-of-Experts (MoE) models on edge devices presents significant challenges due to memory constraints. While MoE architectures enable efficient utilization of computational resources by activating only a subset of experts per inference, they require careful memory management to operate efficiently in resource-constrained environments. Traditional heuristic-based expert caching strategies such as MoE-Infinity struggle to maintain high cache hit rates as models parameters scale. In this work, we introduce MoE-Beyond, a learning-based expert activation predictor "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.17137","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/2508.17137/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":"2508.17137","created_at":"2026-07-05T11:58:10.106537+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.17137v1","created_at":"2026-07-05T11:58:10.106537+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.17137","created_at":"2026-07-05T11:58:10.106537+00:00"},{"alias_kind":"pith_short_12","alias_value":"BZKGOXZUT6FQ","created_at":"2026-07-05T11:58:10.106537+00:00"},{"alias_kind":"pith_short_16","alias_value":"BZKGOXZUT6FQUF7O","created_at":"2026-07-05T11:58:10.106537+00:00"},{"alias_kind":"pith_short_8","alias_value":"BZKGOXZU","created_at":"2026-07-05T11:58:10.106537+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.07911","citing_title":"When Does Trace-Driven Evaluation Mislead MoE Expert Caching? Replay Semantics, Workload Contamination, and Operating Regimes","ref_index":11,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5","json":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5.json","graph_json":"https://pith.science/api/pith-number/BZKGOXZUT6FQUF7OM7X5UOQFW5/graph.json","events_json":"https://pith.science/api/pith-number/BZKGOXZUT6FQUF7OM7X5UOQFW5/events.json","paper":"https://pith.science/paper/BZKGOXZU"},"agent_actions":{"view_html":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5","download_json":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5.json","view_paper":"https://pith.science/paper/BZKGOXZU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.17137&json=true","fetch_graph":"https://pith.science/api/pith-number/BZKGOXZUT6FQUF7OM7X5UOQFW5/graph.json","fetch_events":"https://pith.science/api/pith-number/BZKGOXZUT6FQUF7OM7X5UOQFW5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5/action/storage_attestation","attest_author":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5/action/author_attestation","sign_citation":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5/action/citation_signature","submit_replication":"https://pith.science/pith/BZKGOXZUT6FQUF7OM7X5UOQFW5/action/replication_record"}},"created_at":"2026-07-05T11:58:10.106537+00:00","updated_at":"2026-07-05T11:58:10.106537+00:00"}