{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Q6PZTPQB23TNHG5IVNDCTSMRHX","short_pith_number":"pith:Q6PZTPQB","schema_version":"1.0","canonical_sha256":"879f99be01d6e6d39ba8ab4629c9913dd489a0de349a583baccf10e2d94d4a3c","source":{"kind":"arxiv","id":"2505.17553","version":2},"attestation_state":"computed","paper":{"title":"CoMoE: Contrastive Representation for Mixture-of-Experts in Parameter-Efficient Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Chaopeng Wei, Jinyuan Feng, Tenghai Qiu, Tianyi Hu, Zhiqiang Pu","submitted_at":"2025-05-23T06:58:44Z","abstract_excerpt":"In parameter-efficient fine-tuning, mixture-of-experts (MoE), which involves specializing functionalities into different experts and sparsely activating them appropriately, has been widely adopted as a promising approach to trade-off between model capacity and computation overhead. However, current MoE variants fall short on heterogeneous datasets, ignoring the fact that experts may learn similar knowledge, resulting in the underutilization of MoE's capacity. In this paper, we propose Contrastive Representation for MoE (CoMoE), a novel method to promote modularization and specialization in MoE"},"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":"2505.17553","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-23T06:58:44Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"1dd6c02922e01782f830e420072850109787415d6022203675d05359f4a46b86","abstract_canon_sha256":"6dc1afc2fe88809cdd4a8df9dfa4e48d988ee5465e7692e666a98aa7106de333"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:00:50.090796Z","signature_b64":"5P9LESLhDYhS4ihUcTa4KPOhFWCc40syGpUd8806a0tfQVY1+EWnHWb5EJPBEKeSLzO2PKoiaIQT7yvShy6JAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"879f99be01d6e6d39ba8ab4629c9913dd489a0de349a583baccf10e2d94d4a3c","last_reissued_at":"2026-07-05T12:00:50.090320Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:00:50.090320Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CoMoE: Contrastive Representation for Mixture-of-Experts in Parameter-Efficient Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Chaopeng Wei, Jinyuan Feng, Tenghai Qiu, Tianyi Hu, Zhiqiang Pu","submitted_at":"2025-05-23T06:58:44Z","abstract_excerpt":"In parameter-efficient fine-tuning, mixture-of-experts (MoE), which involves specializing functionalities into different experts and sparsely activating them appropriately, has been widely adopted as a promising approach to trade-off between model capacity and computation overhead. However, current MoE variants fall short on heterogeneous datasets, ignoring the fact that experts may learn similar knowledge, resulting in the underutilization of MoE's capacity. In this paper, we propose Contrastive Representation for MoE (CoMoE), a novel method to promote modularization and specialization in MoE"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.17553","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/2505.17553/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":"2505.17553","created_at":"2026-07-05T12:00:50.090380+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.17553v2","created_at":"2026-07-05T12:00:50.090380+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.17553","created_at":"2026-07-05T12:00:50.090380+00:00"},{"alias_kind":"pith_short_12","alias_value":"Q6PZTPQB23TN","created_at":"2026-07-05T12:00:50.090380+00:00"},{"alias_kind":"pith_short_16","alias_value":"Q6PZTPQB23TNHG5I","created_at":"2026-07-05T12:00:50.090380+00:00"},{"alias_kind":"pith_short_8","alias_value":"Q6PZTPQB","created_at":"2026-07-05T12:00:50.090380+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29425","citing_title":"Mixture of Debaters: Learn to Debate at Architectural Level in Multi-Agent Reasoning","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29613","citing_title":"Does Role Specialization Matter for Explanation Faithfulness in Mixture-of-Experts?","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23996","citing_title":"SMoES: Soft Modality-Guided Expert Specialization in MoE-VLMs","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX","json":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX.json","graph_json":"https://pith.science/api/pith-number/Q6PZTPQB23TNHG5IVNDCTSMRHX/graph.json","events_json":"https://pith.science/api/pith-number/Q6PZTPQB23TNHG5IVNDCTSMRHX/events.json","paper":"https://pith.science/paper/Q6PZTPQB"},"agent_actions":{"view_html":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX","download_json":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX.json","view_paper":"https://pith.science/paper/Q6PZTPQB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.17553&json=true","fetch_graph":"https://pith.science/api/pith-number/Q6PZTPQB23TNHG5IVNDCTSMRHX/graph.json","fetch_events":"https://pith.science/api/pith-number/Q6PZTPQB23TNHG5IVNDCTSMRHX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX/action/storage_attestation","attest_author":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX/action/author_attestation","sign_citation":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX/action/citation_signature","submit_replication":"https://pith.science/pith/Q6PZTPQB23TNHG5IVNDCTSMRHX/action/replication_record"}},"created_at":"2026-07-05T12:00:50.090380+00:00","updated_at":"2026-07-05T12:00:50.090380+00:00"}