{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KYFG42MW7WELU5NGHCXFQ5HV63","short_pith_number":"pith:KYFG42MW","schema_version":"1.0","canonical_sha256":"560a6e6996fd88ba75a638ae5874f5f6ea1c2f4be251fc01b1b3db4e8db076d6","source":{"kind":"arxiv","id":"2505.13273","version":2},"attestation_state":"computed","paper":{"title":"EMoE: Training-Free Expert Disagreement for Uncertainty-Aware Text-to-Image Diffusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Axel Brando, David Meger, Juan Camilo Gamboa Higuera, Lucas Berry, Wei-Di Chang","submitted_at":"2025-05-19T15:53:32Z","abstract_excerpt":"Large text-to-image diffusion models rarely expose reliable signals of when a prompt is likely to produce a poorly aligned generation, especially when training data is undisclosed. We study whether expert disagreement inside pre-trained mixture-of-experts (MoE) diffusion models can serve as a reliable estimate for epistemic uncertainty. We introduce EMoE, a training-free method that separates expert-specific computation paths at an early MoE layer, uses the same initial noise across paths, and measures variance among their latent representations after the first denoising step. This provides 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":"2505.13273","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-19T15:53:32Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"305c02324cd848b5e6c60068d8fe120cdb6b28bd028ea1b9dfe1135c7e55cafb","abstract_canon_sha256":"86ae42f5e13fd676cbbdf454870fb0e5e9096f70b28ffafa53d506ebbc0f67cf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:14.943576Z","signature_b64":"6HeHwoh3fQyc41kVPq+d/3BH/IqJGSiKk7lns0G5yPfhr7k1kGI6yUTsn9hYu2mO1ouoydYrlMlVZSZOjQ6UCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"560a6e6996fd88ba75a638ae5874f5f6ea1c2f4be251fc01b1b3db4e8db076d6","last_reissued_at":"2026-06-02T01:03:14.943094Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:14.943094Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EMoE: Training-Free Expert Disagreement for Uncertainty-Aware Text-to-Image Diffusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Axel Brando, David Meger, Juan Camilo Gamboa Higuera, Lucas Berry, Wei-Di Chang","submitted_at":"2025-05-19T15:53:32Z","abstract_excerpt":"Large text-to-image diffusion models rarely expose reliable signals of when a prompt is likely to produce a poorly aligned generation, especially when training data is undisclosed. We study whether expert disagreement inside pre-trained mixture-of-experts (MoE) diffusion models can serve as a reliable estimate for epistemic uncertainty. We introduce EMoE, a training-free method that separates expert-specific computation paths at an early MoE layer, uses the same initial noise across paths, and measures variance among their latent representations after the first denoising step. This provides an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.13273","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/2505.13273/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":"2505.13273","created_at":"2026-06-02T01:03:14.943152+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.13273v2","created_at":"2026-06-02T01:03:14.943152+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.13273","created_at":"2026-06-02T01:03:14.943152+00:00"},{"alias_kind":"pith_short_12","alias_value":"KYFG42MW7WEL","created_at":"2026-06-02T01:03:14.943152+00:00"},{"alias_kind":"pith_short_16","alias_value":"KYFG42MW7WELU5NG","created_at":"2026-06-02T01:03:14.943152+00:00"},{"alias_kind":"pith_short_8","alias_value":"KYFG42MW","created_at":"2026-06-02T01:03:14.943152+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/KYFG42MW7WELU5NGHCXFQ5HV63","json":"https://pith.science/pith/KYFG42MW7WELU5NGHCXFQ5HV63.json","graph_json":"https://pith.science/api/pith-number/KYFG42MW7WELU5NGHCXFQ5HV63/graph.json","events_json":"https://pith.science/api/pith-number/KYFG42MW7WELU5NGHCXFQ5HV63/events.json","paper":"https://pith.science/paper/KYFG42MW"},"agent_actions":{"view_html":"https://pith.science/pith/KYFG42MW7WELU5NGHCXFQ5HV63","download_json":"https://pith.science/pith/KYFG42MW7WELU5NGHCXFQ5HV63.json","view_paper":"https://pith.science/paper/KYFG42MW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.13273&json=true","fetch_graph":"https://pith.science/api/pith-number/KYFG42MW7WELU5NGHCXFQ5HV63/graph.json","fetch_events":"https://pith.science/api/pith-number/KYFG42MW7WELU5NGHCXFQ5HV63/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KYFG42MW7WELU5NGHCXFQ5HV63/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KYFG42MW7WELU5NGHCXFQ5HV63/action/storage_attestation","attest_author":"https://pith.science/pith/KYFG42MW7WELU5NGHCXFQ5HV63/action/author_attestation","sign_citation":"https://pith.science/pith/KYFG42MW7WELU5NGHCXFQ5HV63/action/citation_signature","submit_replication":"https://pith.science/pith/KYFG42MW7WELU5NGHCXFQ5HV63/action/replication_record"}},"created_at":"2026-06-02T01:03:14.943152+00:00","updated_at":"2026-06-02T01:03:14.943152+00:00"}