{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:X5HLR35ABEVRWIQTMZXKSPXDDP","short_pith_number":"pith:X5HLR35A","schema_version":"1.0","canonical_sha256":"bf4eb8efa0092b1b2213666ea93ee31bfccd499839a05100a25ddcfc49de06ca","source":{"kind":"arxiv","id":"2606.01509","version":1},"attestation_state":"computed","paper":{"title":"ProbMoE: Differentiable Probabilistic Routing for Mixture-of-Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Guy Van den Broeck, Heng Zhao, Zhe Zeng, Zilei Shao","submitted_at":"2026-06-01T00:16:36Z","abstract_excerpt":"Mixture-of-Experts (MoE) models scale by activating only a small subset of experts per token. However, training such models remains challenging because top-$k$ routing is discrete and non-differentiable, requiring gradient estimators for expert selection whose design remains a central open problem. We introduce ProbMoE, a probabilistic routing framework that models expert selection as a distribution over cardinality-constrained expert subsets and formulates routing as probabilistic inference in this discrete subset space. We first propose ProbMoE Exact-$k$ routing, which samples $k$-expert sub"},"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":"2606.01509","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-01T00:16:36Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"88aeb6fa964d069d2bc9f05afa8899599b0d7e80167d2e290de84dda87959f53","abstract_canon_sha256":"04e718cf8aad4aabbd48a70667aae9954c0f77c131e1d1db36a5d622d3add7e3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:35.059039Z","signature_b64":"HySxDMq5rg/NkHXpIhBMtZYW5On0lChV69rb06u3n5aC628T3m0sv63ZF7LuO+OHPpKFov7SwE016RXnf9KLAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bf4eb8efa0092b1b2213666ea93ee31bfccd499839a05100a25ddcfc49de06ca","last_reissued_at":"2026-06-02T02:04:35.058524Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:35.058524Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ProbMoE: Differentiable Probabilistic Routing for Mixture-of-Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Guy Van den Broeck, Heng Zhao, Zhe Zeng, Zilei Shao","submitted_at":"2026-06-01T00:16:36Z","abstract_excerpt":"Mixture-of-Experts (MoE) models scale by activating only a small subset of experts per token. However, training such models remains challenging because top-$k$ routing is discrete and non-differentiable, requiring gradient estimators for expert selection whose design remains a central open problem. We introduce ProbMoE, a probabilistic routing framework that models expert selection as a distribution over cardinality-constrained expert subsets and formulates routing as probabilistic inference in this discrete subset space. We first propose ProbMoE Exact-$k$ routing, which samples $k$-expert sub"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.01509","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/2606.01509/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":"2606.01509","created_at":"2026-06-02T02:04:35.058595+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.01509v1","created_at":"2026-06-02T02:04:35.058595+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.01509","created_at":"2026-06-02T02:04:35.058595+00:00"},{"alias_kind":"pith_short_12","alias_value":"X5HLR35ABEVR","created_at":"2026-06-02T02:04:35.058595+00:00"},{"alias_kind":"pith_short_16","alias_value":"X5HLR35ABEVRWIQT","created_at":"2026-06-02T02:04:35.058595+00:00"},{"alias_kind":"pith_short_8","alias_value":"X5HLR35A","created_at":"2026-06-02T02:04:35.058595+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/X5HLR35ABEVRWIQTMZXKSPXDDP","json":"https://pith.science/pith/X5HLR35ABEVRWIQTMZXKSPXDDP.json","graph_json":"https://pith.science/api/pith-number/X5HLR35ABEVRWIQTMZXKSPXDDP/graph.json","events_json":"https://pith.science/api/pith-number/X5HLR35ABEVRWIQTMZXKSPXDDP/events.json","paper":"https://pith.science/paper/X5HLR35A"},"agent_actions":{"view_html":"https://pith.science/pith/X5HLR35ABEVRWIQTMZXKSPXDDP","download_json":"https://pith.science/pith/X5HLR35ABEVRWIQTMZXKSPXDDP.json","view_paper":"https://pith.science/paper/X5HLR35A","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.01509&json=true","fetch_graph":"https://pith.science/api/pith-number/X5HLR35ABEVRWIQTMZXKSPXDDP/graph.json","fetch_events":"https://pith.science/api/pith-number/X5HLR35ABEVRWIQTMZXKSPXDDP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X5HLR35ABEVRWIQTMZXKSPXDDP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X5HLR35ABEVRWIQTMZXKSPXDDP/action/storage_attestation","attest_author":"https://pith.science/pith/X5HLR35ABEVRWIQTMZXKSPXDDP/action/author_attestation","sign_citation":"https://pith.science/pith/X5HLR35ABEVRWIQTMZXKSPXDDP/action/citation_signature","submit_replication":"https://pith.science/pith/X5HLR35ABEVRWIQTMZXKSPXDDP/action/replication_record"}},"created_at":"2026-06-02T02:04:35.058595+00:00","updated_at":"2026-06-02T02:04:35.058595+00:00"}