{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:OZUPR4BVTWDMYSSOE6XPJ2ALWW","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"4fc87225d0bf537713fef270d3fb111cfc24792dc12e6e3522fcc78d008b868f","cross_cats_sorted":["math.OC","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T20:39:28Z","title_canon_sha256":"3562498ca70109c3ba26eca3a5d2c962940053ae63d5d47f750bc534399aff91"},"schema_version":"1.0","source":{"id":"2605.15403","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15403","created_at":"2026-05-20T00:00:56Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15403v1","created_at":"2026-05-20T00:00:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15403","created_at":"2026-05-20T00:00:56Z"},{"alias_kind":"pith_short_12","alias_value":"OZUPR4BVTWDM","created_at":"2026-05-20T00:00:56Z"},{"alias_kind":"pith_short_16","alias_value":"OZUPR4BVTWDMYSSO","created_at":"2026-05-20T00:00:56Z"},{"alias_kind":"pith_short_8","alias_value":"OZUPR4BV","created_at":"2026-05-20T00:00:56Z"}],"graph_snapshots":[{"event_id":"sha256:71a1a86a8d068eb9b766cf3cc913ac95b4b6e18c9fa6a24d6216d15fb70e444d","target":"graph","created_at":"2026-05-20T00:00:56Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Across large-scale pretraining and downstream fine-tuning, φ-balancing consistently outperforms prior Switch-style and loss-free baselines, demonstrating more stable and effective expert utilization."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That minimizing the chosen strictly convex potential of the expected routing distribution produces the desired population-level balance and that the EMA-based online approximation via mirror descent faithfully tracks the population objective without introducing new bias (abstract, paragraph on derivation)."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"φ-balancing is a convex optimization method for population-level expert balance in MoE training that derives an online EMA adjustment and outperforms heuristic baselines."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Mixture-of-experts models achieve population-level expert balance by minimizing a strictly convex potential of the expected routing distribution."}],"snapshot_sha256":"955ebbb2ec6dbf24025c7adfc1f83c49aa3ea386480fab272d414054421ae24d"},"formal_canon":{"evidence_count":3,"snapshot_sha256":"aecd9437594667a2038f9826e20a13b2388264e66509285569391283b61e13ad"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T16:31:18.268850Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T16:25:54.221458Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"cited_work_retraction","ran_at":"2026-05-19T16:23:51.333105Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"citation_quote_validity","ran_at":"2026-05-19T15:50:50.932555Z","status":"skipped","version":"0.1.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T14:21:54.158666Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.715382Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.15403/integrity.json","findings":[],"snapshot_sha256":"ac86f12ecb5a351f2b55165686c2a843bd71bf0d5f5c8899befc32acb3aaf78f","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Mixture-of-Experts (MoE) models rely on balanced expert utilization to fully realize their scalability. However, existing load-balancing methods are largely heuristic and operate on noisy mini-batch assignment statistics, introducing bias relative to population-level objectives. We propose $\\phi$-balancing, a principled framework that directly targets population-level expert balance by minimizing a strictly convex, symmetric, and differentiable potential of the expected routing distribution. Using convex duality, we derive an equivalent min-max formulation and obtain a simple online algorithm ","authors_text":"Chen Liang, Jonathan Li, Lizhang Chen, Ni Lao, Qiang Liu, Qi Wang, Runlong Liao, Shuozhe Li","cross_cats":["math.OC","stat.ML"],"headline":"Mixture-of-experts models achieve population-level expert balance by minimizing a strictly convex potential of the expected routing distribution.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T20:39:28Z","title":"$\\phi$-Balancing for Mixture-of-Experts Training"},"references":{"count":49,"internal_anchors":9,"resolved_work":49,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Lion secretly solves a constrained optimization: As Lyapunov predicts","work_id":"86bf0e47-3d82-47e0-ad43-dbf2d3122d34","year":2024},{"cited_arxiv_id":"2107.03374","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Chen, X., Liang, C., Huang, D., Real, E., Wang, K., Pham, H., Dong, X., Luong, T., Hsieh, C., Lu, Y ., and Le, Q. V . Symbolic discovery of optimization algorithms. InAd- vances in Neural Information ","work_id":"57bc6503-bb9d-47f9-a122-93d5ebec65df","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Ermoe: Eigen- reparameterized mixture-of-experts for stable routing and interpretable specialization","work_id":"61ec3cfe-6814-43fb-92f6-ccf0b2069255","year":null},{"cited_arxiv_id":"2110.14168","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","year":null}],"snapshot_sha256":"d244ef97aa28f26cd554192144bf0a2ad7cf80982edb51eabf24ffab140ada16"},"source":{"id":"2605.15403","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T16:11:46.193593Z","id":"0a675f24-6592-4b5c-8477-cb87f4cc71cd","model_set":{"reader":"grok-4.3"},"one_line_summary":"φ-balancing is a convex optimization method for population-level expert balance in MoE training that derives an online EMA adjustment and outperforms heuristic baselines.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Mixture-of-experts models achieve population-level expert balance by minimizing a strictly convex potential of the expected routing distribution.","strongest_claim":"Across large-scale pretraining and downstream fine-tuning, φ-balancing consistently outperforms prior Switch-style and loss-free baselines, demonstrating more stable and effective expert utilization.","weakest_assumption":"That minimizing the chosen strictly convex potential of the expected routing distribution produces the desired population-level balance and that the EMA-based online approximation via mirror descent faithfully tracks the population objective without introducing new bias (abstract, paragraph on derivation)."}},"verdict_id":"0a675f24-6592-4b5c-8477-cb87f4cc71cd"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:a6f25712010081b3677065313f1d4b1af798cb18e355a2005533c438b92b49ef","target":"record","created_at":"2026-05-20T00:00:56Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"4fc87225d0bf537713fef270d3fb111cfc24792dc12e6e3522fcc78d008b868f","cross_cats_sorted":["math.OC","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T20:39:28Z","title_canon_sha256":"3562498ca70109c3ba26eca3a5d2c962940053ae63d5d47f750bc534399aff91"},"schema_version":"1.0","source":{"id":"2605.15403","kind":"arxiv","version":1}},"canonical_sha256":"7668f8f0359d86cc4a4e27aef4e80bb5ba698b537c1265cabd5e7de8bbcb095f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7668f8f0359d86cc4a4e27aef4e80bb5ba698b537c1265cabd5e7de8bbcb095f","first_computed_at":"2026-05-20T00:00:56.875852Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:00:56.875852Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"R8yxPwQCZwquD8o28JCenf9AxPOXdQyEk+xViM5qwupntdYEgh7X6P6D+TRiriN/uabKp80TlaaNwNB7tAERDA==","signature_status":"signed_v1","signed_at":"2026-05-20T00:00:56.876483Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.15403","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a6f25712010081b3677065313f1d4b1af798cb18e355a2005533c438b92b49ef","sha256:71a1a86a8d068eb9b766cf3cc913ac95b4b6e18c9fa6a24d6216d15fb70e444d"],"state_sha256":"440ee0b2b95370bfd3a3efe037ad291ecc4529e0e65da20a667af93852db2b5a"}