{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BNPOUUG6PCKXTKSVLVPZM3R7EQ","short_pith_number":"pith:BNPOUUG6","schema_version":"1.0","canonical_sha256":"0b5eea50de789579aa555d5f966e3f2408c2c7a9c90d81a233e1495dd87e2442","source":{"kind":"arxiv","id":"2503.14487","version":1},"attestation_state":"computed","paper":{"title":"DiffMoE: Dynamic Token Selection for Scalable Diffusion Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Di Zhang, Haotian Yang, Jie Zhou, Jiwen Lu, Kun Gai, Minglei Shi, Mingwu Zheng, Pengfei Wan, Wenliang Zhao, Wenzhao Zheng, Xin Tao, Xintao Wang, Ziyang Yuan","submitted_at":"2025-03-18T17:57:07Z","abstract_excerpt":"Diffusion models have demonstrated remarkable success in various image generation tasks, but their performance is often limited by the uniform processing of inputs across varying conditions and noise levels. To address this limitation, we propose a novel approach that leverages the inherent heterogeneity of the diffusion process. Our method, DiffMoE, introduces a batch-level global token pool that enables experts to access global token distributions during training, promoting specialized expert behavior. To unleash the full potential of the diffusion process, DiffMoE incorporates a capacity pr"},"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":"2503.14487","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-03-18T17:57:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"812c30ad088a88fffde78fdb044ad7090524fcb1ac94fdf6be0baf83ca4fdef2","abstract_canon_sha256":"105ececd66cd2487f893d533aad75cb36be542e2fdc6c2ccfba509e2370da7e0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:34:03.883971Z","signature_b64":"P+lZ9sW4ra5UKg5Nk1MQiET4fA6aRtQfvuoth85vgQ5nQzE7Sl8QBSSgqlO09RFWQvZYGMyfLiKTeemX/31jCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0b5eea50de789579aa555d5f966e3f2408c2c7a9c90d81a233e1495dd87e2442","last_reissued_at":"2026-07-05T10:34:03.883309Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:34:03.883309Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DiffMoE: Dynamic Token Selection for Scalable Diffusion Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Di Zhang, Haotian Yang, Jie Zhou, Jiwen Lu, Kun Gai, Minglei Shi, Mingwu Zheng, Pengfei Wan, Wenliang Zhao, Wenzhao Zheng, Xin Tao, Xintao Wang, Ziyang Yuan","submitted_at":"2025-03-18T17:57:07Z","abstract_excerpt":"Diffusion models have demonstrated remarkable success in various image generation tasks, but their performance is often limited by the uniform processing of inputs across varying conditions and noise levels. To address this limitation, we propose a novel approach that leverages the inherent heterogeneity of the diffusion process. Our method, DiffMoE, introduces a batch-level global token pool that enables experts to access global token distributions during training, promoting specialized expert behavior. To unleash the full potential of the diffusion process, DiffMoE incorporates a capacity pr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.14487","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/2503.14487/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":"2503.14487","created_at":"2026-07-05T10:34:03.883387+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.14487v1","created_at":"2026-07-05T10:34:03.883387+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.14487","created_at":"2026-07-05T10:34:03.883387+00:00"},{"alias_kind":"pith_short_12","alias_value":"BNPOUUG6PCKX","created_at":"2026-07-05T10:34:03.883387+00:00"},{"alias_kind":"pith_short_16","alias_value":"BNPOUUG6PCKXTKSV","created_at":"2026-07-05T10:34:03.883387+00:00"},{"alias_kind":"pith_short_8","alias_value":"BNPOUUG6","created_at":"2026-07-05T10:34:03.883387+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26938","citing_title":"Focusing on What Matters: Saliency-Harnessing Accurate Routing for Diffusion MoE","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21033","citing_title":"MoECodec: Image Compression for joint human and machine perception via Mixture-of-Experts","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30045","citing_title":"GenEraser: Generalizable Video Object Removal via Balanced Text-Mask Guidance and Decoupled Locator-Preserver","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2603.06351","citing_title":"DC-DiT: Adaptive Compute and Elastic Inference for Visual Generation via Dynamic Chunking","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19636","citing_title":"CoInteract: Physically-Consistent Human-Object Interaction Video Synthesis via Spatially-Structured Co-Generation","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02641","citing_title":"Mamoda2.5: Enhancing Unified Multimodal Model with DiT-MoE","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ","json":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ.json","graph_json":"https://pith.science/api/pith-number/BNPOUUG6PCKXTKSVLVPZM3R7EQ/graph.json","events_json":"https://pith.science/api/pith-number/BNPOUUG6PCKXTKSVLVPZM3R7EQ/events.json","paper":"https://pith.science/paper/BNPOUUG6"},"agent_actions":{"view_html":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ","download_json":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ.json","view_paper":"https://pith.science/paper/BNPOUUG6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.14487&json=true","fetch_graph":"https://pith.science/api/pith-number/BNPOUUG6PCKXTKSVLVPZM3R7EQ/graph.json","fetch_events":"https://pith.science/api/pith-number/BNPOUUG6PCKXTKSVLVPZM3R7EQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ/action/storage_attestation","attest_author":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ/action/author_attestation","sign_citation":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ/action/citation_signature","submit_replication":"https://pith.science/pith/BNPOUUG6PCKXTKSVLVPZM3R7EQ/action/replication_record"}},"created_at":"2026-07-05T10:34:03.883387+00:00","updated_at":"2026-07-05T10:34:03.883387+00:00"}