{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:OIWPZEJENDCTQ77VFMLTOXUJ6B","short_pith_number":"pith:OIWPZEJE","schema_version":"1.0","canonical_sha256":"722cfc912468c5387ff52b17375e89f0497882eb8719ca404ba39745cb4bcdc5","source":{"kind":"arxiv","id":"2505.00792","version":1},"attestation_state":"computed","paper":{"title":"Improving Routing in Sparse Mixture of Experts with Graph of Tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Khai Nguyen, Ngoc N. Tran, Richard G. Baraniuk, Tam Nguyen","submitted_at":"2025-05-01T18:44:20Z","abstract_excerpt":"Sparse Mixture of Experts (SMoE) has emerged as a key to achieving unprecedented scalability in deep learning. By activating only a small subset of parameters per sample, SMoE achieves an exponential increase in parameter counts while maintaining a constant computational overhead. However, SMoE models are susceptible to routing fluctuations--changes in the routing of a given input to its target expert--at the late stage of model training, leading to model non-robustness. In this work, we unveil the limitation of SMoE through the perspective of the probabilistic graphical model (PGM). Through t"},"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":"2505.00792","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-01T18:44:20Z","cross_cats_sorted":[],"title_canon_sha256":"dcd956114378c8d5130f8918841d990c78c8a2c1d6ac61e62ff9df173d26c715","abstract_canon_sha256":"f7cf5b046309bf63e6056a7d8bc4d4d59fff7d0f8b08bf62322ffed98ca7b23a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:57:39.122602Z","signature_b64":"iaXZ+8NczPnM9Hlj/EqBjO4pLw8amWT1YFkTaJUdglreEVCymqzVQU+rDDPBKUYUhk//RyYUwSzlzvRwIqIECg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"722cfc912468c5387ff52b17375e89f0497882eb8719ca404ba39745cb4bcdc5","last_reissued_at":"2026-07-05T10:57:39.122057Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:57:39.122057Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Routing in Sparse Mixture of Experts with Graph of Tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Khai Nguyen, Ngoc N. Tran, Richard G. Baraniuk, Tam Nguyen","submitted_at":"2025-05-01T18:44:20Z","abstract_excerpt":"Sparse Mixture of Experts (SMoE) has emerged as a key to achieving unprecedented scalability in deep learning. By activating only a small subset of parameters per sample, SMoE achieves an exponential increase in parameter counts while maintaining a constant computational overhead. However, SMoE models are susceptible to routing fluctuations--changes in the routing of a given input to its target expert--at the late stage of model training, leading to model non-robustness. In this work, we unveil the limitation of SMoE through the perspective of the probabilistic graphical model (PGM). Through t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.00792","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/2505.00792/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":"2505.00792","created_at":"2026-07-05T10:57:39.122116+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.00792v1","created_at":"2026-07-05T10:57:39.122116+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.00792","created_at":"2026-07-05T10:57:39.122116+00:00"},{"alias_kind":"pith_short_12","alias_value":"OIWPZEJENDCT","created_at":"2026-07-05T10:57:39.122116+00:00"},{"alias_kind":"pith_short_16","alias_value":"OIWPZEJENDCTQ77V","created_at":"2026-07-05T10:57:39.122116+00:00"},{"alias_kind":"pith_short_8","alias_value":"OIWPZEJE","created_at":"2026-07-05T10:57:39.122116+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.01047","citing_title":"Learning Multi-Modal Trajectory Policies for Data-Efficient Robotic Manipulation","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07298","citing_title":"Region-Graph Optimal Transport Routing for Mixture-of-Experts Whole-Slide Image Classification","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B","json":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B.json","graph_json":"https://pith.science/api/pith-number/OIWPZEJENDCTQ77VFMLTOXUJ6B/graph.json","events_json":"https://pith.science/api/pith-number/OIWPZEJENDCTQ77VFMLTOXUJ6B/events.json","paper":"https://pith.science/paper/OIWPZEJE"},"agent_actions":{"view_html":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B","download_json":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B.json","view_paper":"https://pith.science/paper/OIWPZEJE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.00792&json=true","fetch_graph":"https://pith.science/api/pith-number/OIWPZEJENDCTQ77VFMLTOXUJ6B/graph.json","fetch_events":"https://pith.science/api/pith-number/OIWPZEJENDCTQ77VFMLTOXUJ6B/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B/action/storage_attestation","attest_author":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B/action/author_attestation","sign_citation":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B/action/citation_signature","submit_replication":"https://pith.science/pith/OIWPZEJENDCTQ77VFMLTOXUJ6B/action/replication_record"}},"created_at":"2026-07-05T10:57:39.122116+00:00","updated_at":"2026-07-05T10:57:39.122116+00:00"}