{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:PIUWM3KXZNLTR7DXRAQIXFE4UX","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":"c026d7764370c9b6d47a32b6759fd64030945d3f46e0be17032249b380dbb6ab","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-07-31T17:46:51Z","title_canon_sha256":"d7adca15f07dfa92b030d089c2da71ffcb3a9feff1a708fd1917fb6329dc009d"},"schema_version":"1.0","source":{"id":"2407.21770","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.21770","created_at":"2026-07-05T08:54:36Z"},{"alias_kind":"arxiv_version","alias_value":"2407.21770v3","created_at":"2026-07-05T08:54:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.21770","created_at":"2026-07-05T08:54:36Z"},{"alias_kind":"pith_short_12","alias_value":"PIUWM3KXZNLT","created_at":"2026-07-05T08:54:36Z"},{"alias_kind":"pith_short_16","alias_value":"PIUWM3KXZNLTR7DX","created_at":"2026-07-05T08:54:36Z"},{"alias_kind":"pith_short_8","alias_value":"PIUWM3KX","created_at":"2026-07-05T08:54:36Z"}],"graph_snapshots":[{"event_id":"sha256:321aa0f88c26630e8f09bba4247721edc0b728a2ab053e48c480eef9678de14f","target":"graph","created_at":"2026-07-05T08:54:36Z","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/2407.21770/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We introduce MoMa, a novel modality-aware mixture-of-experts (MoE) architecture designed for pre-training mixed-modal, early-fusion language models. MoMa processes images and text in arbitrary sequences by dividing expert modules into modality-specific groups. These groups exclusively process designated tokens while employing learned routing within each group to maintain semantically informed adaptivity. Our empirical results reveal substantial pre-training efficiency gains through this modality-specific parameter allocation. Under a 1-trillion-token training budget, the MoMa 1.4B model, featu","authors_text":"Akshat Shrivastava, Armen Aghajanyan, Gargi Ghosh, Liang Luo, Luke Zettlemoyer, Mike Lewis, Srinivasan Iyer, Xi Victoria Lin","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-07-31T17:46:51Z","title":"MoMa: Efficient Early-Fusion Pre-training with Mixture of Modality-Aware Experts"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.21770","kind":"arxiv","version":3},"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":"sha256:7919be4d0d04c812cdf02cbacc68a16b2b71db9b356890ec5cfc34232098da4e","target":"record","created_at":"2026-07-05T08:54:36Z","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":"c026d7764370c9b6d47a32b6759fd64030945d3f46e0be17032249b380dbb6ab","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-07-31T17:46:51Z","title_canon_sha256":"d7adca15f07dfa92b030d089c2da71ffcb3a9feff1a708fd1917fb6329dc009d"},"schema_version":"1.0","source":{"id":"2407.21770","kind":"arxiv","version":3}},"canonical_sha256":"7a29666d57cb5738fc7788208b949ca5d93568645d686db40a56dd1ee5fa7d33","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7a29666d57cb5738fc7788208b949ca5d93568645d686db40a56dd1ee5fa7d33","first_computed_at":"2026-07-05T08:54:36.851005Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:54:36.851005Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YaiczEKW5I5oEHQ+Lz7uz9TSlC8nB7enQVsQKPw2NKQ+FVB64ZGTKvIsYEcyI5sEto9lgaKqe9yGLgr64NLwDA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:54:36.851430Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.21770","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7919be4d0d04c812cdf02cbacc68a16b2b71db9b356890ec5cfc34232098da4e","sha256:321aa0f88c26630e8f09bba4247721edc0b728a2ab053e48c480eef9678de14f"],"state_sha256":"6780b28757538fb9b35e392b3c4715db14ba7620422144743fa7fee9e3e6bc3f"}