{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:AO5CUTAYHSQLUHCARBOGOLKD7H","short_pith_number":"pith:AO5CUTAY","schema_version":"1.0","canonical_sha256":"03ba2a4c183ca0ba1c40885c672d43f9fb260e72e5fe83cf4a90c9a03b5fdafb","source":{"kind":"arxiv","id":"2607.01789","version":1},"attestation_state":"computed","paper":{"title":"EPnG: Adaptive Expert Prune-and-Grow for Parameter-Efficient MoE Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ahin Lee, Sehyun Yun, Taesik Gong","submitted_at":"2026-07-02T07:02:44Z","abstract_excerpt":"Mixture-of-Experts (MoE) models scale efficiently but remain costly to adapt due to redundant experts and uniform parameter allocation. Existing parameter-efficient fine-tuning (PEFT) methods such as LoRA ignore MoE routing dynamics, leading to suboptimal resource use. We propose EPnG, an adaptive prune-and-grow framework that reallocates LoRA capacity based on expert importance derived from router gate probabilities. EPnG prunes under-utilized experts and expands high-importance experts via rank growth with orthogonal initialization, while maintaining a fixed parameter budget. Across OLMoE an"},"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":"2607.01789","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-07-02T07:02:44Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"40c887cac96519b0a25adc8fde72e28083329929b7409ed8d527823083b32c59","abstract_canon_sha256":"fc0d582837ddcca35e092a2bf2bf00f5368933206fc1649d5e54f59328e3f5af"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T01:17:29.918276Z","signature_b64":"Mo3Ctqr0DFVyp229uBJHz1e6VT9ymlnCNy4imNbKyBlVVaq1hYT83sOccfoJnAwNtQM5ZyAH4MpEP8Xp9WlnBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"03ba2a4c183ca0ba1c40885c672d43f9fb260e72e5fe83cf4a90c9a03b5fdafb","last_reissued_at":"2026-07-03T01:17:29.917868Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T01:17:29.917868Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EPnG: Adaptive Expert Prune-and-Grow for Parameter-Efficient MoE Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ahin Lee, Sehyun Yun, Taesik Gong","submitted_at":"2026-07-02T07:02:44Z","abstract_excerpt":"Mixture-of-Experts (MoE) models scale efficiently but remain costly to adapt due to redundant experts and uniform parameter allocation. Existing parameter-efficient fine-tuning (PEFT) methods such as LoRA ignore MoE routing dynamics, leading to suboptimal resource use. We propose EPnG, an adaptive prune-and-grow framework that reallocates LoRA capacity based on expert importance derived from router gate probabilities. EPnG prunes under-utilized experts and expands high-importance experts via rank growth with orthogonal initialization, while maintaining a fixed parameter budget. Across OLMoE an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.01789","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/2607.01789/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":"2607.01789","created_at":"2026-07-03T01:17:29.917918+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.01789v1","created_at":"2026-07-03T01:17:29.917918+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.01789","created_at":"2026-07-03T01:17:29.917918+00:00"},{"alias_kind":"pith_short_12","alias_value":"AO5CUTAYHSQL","created_at":"2026-07-03T01:17:29.917918+00:00"},{"alias_kind":"pith_short_16","alias_value":"AO5CUTAYHSQLUHCA","created_at":"2026-07-03T01:17:29.917918+00:00"},{"alias_kind":"pith_short_8","alias_value":"AO5CUTAY","created_at":"2026-07-03T01:17:29.917918+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H","json":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H.json","graph_json":"https://pith.science/api/pith-number/AO5CUTAYHSQLUHCARBOGOLKD7H/graph.json","events_json":"https://pith.science/api/pith-number/AO5CUTAYHSQLUHCARBOGOLKD7H/events.json","paper":"https://pith.science/paper/AO5CUTAY"},"agent_actions":{"view_html":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H","download_json":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H.json","view_paper":"https://pith.science/paper/AO5CUTAY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.01789&json=true","fetch_graph":"https://pith.science/api/pith-number/AO5CUTAYHSQLUHCARBOGOLKD7H/graph.json","fetch_events":"https://pith.science/api/pith-number/AO5CUTAYHSQLUHCARBOGOLKD7H/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H/action/storage_attestation","attest_author":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H/action/author_attestation","sign_citation":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H/action/citation_signature","submit_replication":"https://pith.science/pith/AO5CUTAYHSQLUHCARBOGOLKD7H/action/replication_record"}},"created_at":"2026-07-03T01:17:29.917918+00:00","updated_at":"2026-07-03T01:17:29.917918+00:00"}