{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:QGDCLHVKAOJ2GZWWI6JSKPNHQC","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":"05c8173c79994522132a85d809bcfff97cbf3d487d66c08756024ed5251e6bd3","cross_cats_sorted":["cs.CL","cs.CV"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T08:21:28Z","title_canon_sha256":"9cde6dfa693720e15081176a722be32969a36514fa4868f9ad148d30da7c8f60"},"schema_version":"1.0","source":{"id":"2505.17613","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.17613","created_at":"2026-07-05T11:08:28Z"},{"alias_kind":"arxiv_version","alias_value":"2505.17613v1","created_at":"2026-07-05T11:08:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.17613","created_at":"2026-07-05T11:08:28Z"},{"alias_kind":"pith_short_12","alias_value":"QGDCLHVKAOJ2","created_at":"2026-07-05T11:08:28Z"},{"alias_kind":"pith_short_16","alias_value":"QGDCLHVKAOJ2GZWW","created_at":"2026-07-05T11:08:28Z"},{"alias_kind":"pith_short_8","alias_value":"QGDCLHVK","created_at":"2026-07-05T11:08:28Z"}],"graph_snapshots":[{"event_id":"sha256:f7e1ece8c5e089a56c1bf28a71651240dd944818ad3ffd3068849bcc0d9fe39b","target":"graph","created_at":"2026-07-05T11:08:28Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2505.17613/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Automatically evaluating multimodal generation presents a significant challenge, as automated metrics often struggle to align reliably with human evaluation, especially for complex tasks that involve multiple modalities. To address this, we present MMMG, a comprehensive and human-aligned benchmark for multimodal generation across 4 modality combinations (image, audio, interleaved text and image, interleaved text and audio), with a focus on tasks that present significant challenges for generation models, while still enabling reliable automatic evaluation through a combination of models and prog","authors_text":"Banghua Zhu, Bingbing Wen, Bin Han, Guang Yang, Jihan Yao, Lucy Lu Wang, Noah A. Smith, Ranjay Krishna, Shangbin Feng, Yujie Yi, Yulia Tsvetkov, Yushi Hu","cross_cats":["cs.CL","cs.CV"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T08:21:28Z","title":"MMMG: a Comprehensive and Reliable Evaluation Suite for Multitask Multimodal Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.17613","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:0440818dcc958e8a9576d21a124e061f689ca672613b59f296aac2fd14c696fa","target":"record","created_at":"2026-07-05T11:08:28Z","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":"05c8173c79994522132a85d809bcfff97cbf3d487d66c08756024ed5251e6bd3","cross_cats_sorted":["cs.CL","cs.CV"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T08:21:28Z","title_canon_sha256":"9cde6dfa693720e15081176a722be32969a36514fa4868f9ad148d30da7c8f60"},"schema_version":"1.0","source":{"id":"2505.17613","kind":"arxiv","version":1}},"canonical_sha256":"8186259eaa0393a366d64793253da780b34f7fab29f75a4a2a15676f903a6cd6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8186259eaa0393a366d64793253da780b34f7fab29f75a4a2a15676f903a6cd6","first_computed_at":"2026-07-05T11:08:28.993070Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:08:28.993070Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"NDTv0nH4ho5UgqM+nlBhPznZZr+mgtqYDUd7l8+UR9ze41eyOweYu9YC1SLRQadVXtUYSnrH88bbXh5To9dPAA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:08:28.993549Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.17613","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0440818dcc958e8a9576d21a124e061f689ca672613b59f296aac2fd14c696fa","sha256:f7e1ece8c5e089a56c1bf28a71651240dd944818ad3ffd3068849bcc0d9fe39b"],"state_sha256":"02553b1a6fd62abf13df2c858d3d322b3679d06d17e4dd728c0685c56d86786e"}