{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:Y5HR3MRXMK6FSX6PMAM4S5Z6Q4","short_pith_number":"pith:Y5HR3MRX","schema_version":"1.0","canonical_sha256":"c74f1db23762bc595fcf6019c9773e8727f464891dc08adab64526d371aebe71","source":{"kind":"arxiv","id":"2211.01324","version":5},"attestation_state":"computed","paper":{"title":"eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoisers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"An ensemble of stage-specialized diffusion models improves text alignment in image synthesis at the same inference cost.","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Arash Vahdat, Bryan Catanzaro, Jiaming Song, Karsten Kreis, Miika Aittala, Ming-Yu Liu, Qinsheng Zhang, Samuli Laine, Seungjun Nah, Tero Karras, Timo Aila, Xun Huang, Yogesh Balaji","submitted_at":"2022-11-02T17:43:04Z","abstract_excerpt":"Large-scale diffusion-based generative models have led to breakthroughs in text-conditioned high-resolution image synthesis. Starting from random noise, such text-to-image diffusion models gradually synthesize images in an iterative fashion while conditioning on text prompts. We find that their synthesis behavior qualitatively changes throughout this process: Early in sampling, generation strongly relies on the text prompt to generate text-aligned content, while later, the text conditioning is almost entirely ignored. This suggests that sharing model parameters throughout the entire generation"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2211.01324","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2022-11-02T17:43:04Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e9f8d1e62e63e995eafd74d40a92cab36ae1e21fca1d6f3683a0e8e5ac675eea","abstract_canon_sha256":"6ba932880558e507ebb983d38e510f4e566f11bba98dcd82cf7a99666372929f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:05.096476Z","signature_b64":"IjBzyxBOwfyFIG43mI/vCHaKLPtENn6mPd6ux2lYUUJ0K5R57wF5bF9IR3+aJ93AU7KqfBJJOe/H95rkfmtHBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c74f1db23762bc595fcf6019c9773e8727f464891dc08adab64526d371aebe71","last_reissued_at":"2026-05-17T23:39:05.095621Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:05.095621Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoisers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"An ensemble of stage-specialized diffusion models improves text alignment in image synthesis at the same inference cost.","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Arash Vahdat, Bryan Catanzaro, Jiaming Song, Karsten Kreis, Miika Aittala, Ming-Yu Liu, Qinsheng Zhang, Samuli Laine, Seungjun Nah, Tero Karras, Timo Aila, Xun Huang, Yogesh Balaji","submitted_at":"2022-11-02T17:43:04Z","abstract_excerpt":"Large-scale diffusion-based generative models have led to breakthroughs in text-conditioned high-resolution image synthesis. Starting from random noise, such text-to-image diffusion models gradually synthesize images in an iterative fashion while conditioning on text prompts. We find that their synthesis behavior qualitatively changes throughout this process: Early in sampling, generation strongly relies on the text prompt to generate text-aligned content, while later, the text conditioning is almost entirely ignored. This suggests that sharing model parameters throughout the entire generation"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our ensemble of diffusion models, called eDiff-I, results in improved text alignment while maintaining the same inference computation cost and preserving high visual quality, outperforming previous large-scale text-to-image diffusion models on the standard benchmark.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The synthesis behavior qualitatively changes throughout the generation process such that early stages rely on text conditioning while later stages largely ignore it, making a single shared-parameter model suboptimal.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"An ensemble of stage-specialized text-to-image diffusion models improves prompt alignment over single shared-parameter models while preserving visual quality and inference speed.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"An ensemble of stage-specialized diffusion models improves text alignment in image synthesis at the same inference cost.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d19838852ca4de31ef0d83cb2cbcd3555be749baa7f1b69390136d0a2a8993bb"},"source":{"id":"2211.01324","kind":"arxiv","version":5},"verdict":{"id":"ea0e3657-7f80-4a78-848a-196756727a3f","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T01:40:27.202143Z","strongest_claim":"Our ensemble of diffusion models, called eDiff-I, results in improved text alignment while maintaining the same inference computation cost and preserving high visual quality, outperforming previous large-scale text-to-image diffusion models on the standard benchmark.","one_line_summary":"An ensemble of stage-specialized text-to-image diffusion models improves prompt alignment over single shared-parameter models while preserving visual quality and inference speed.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The synthesis behavior qualitatively changes throughout the generation process such that early stages rely on text conditioning while later stages largely ignore it, making a single shared-parameter model suboptimal.","pith_extraction_headline":"An ensemble of stage-specialized diffusion models improves text alignment in image synthesis at the same inference cost."},"references":{"count":95,"sample":[{"doi":"","year":2021,"title":"V ., Du, J., Iyer, S., Pasunuru, R., et al","work_id":"711d98e0-b9e1-4350-b81d-a621de85a6bb","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Blended latent diffusion","work_id":"1aef642e-c5e8-42d8-a391-f41998245355","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Blended diffusion for text-driven editing of natural images","work_id":"b7d69bd8-2483-45c0-93f6-0bf59101b935","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Estimating the optimal covariance with imperfect mean in diffusion probabilistic models","work_id":"6a02cf87-87e3-4e01-8807-1815408583fe","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Analytic- DPM: An analytic estimate of the optimal reverse variance in diffusion probabilistic models","work_id":"83033437-6996-4dfd-9f0f-41f417388b99","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":95,"snapshot_sha256":"534b6679319b65d105d3f306b7c44eee8a40b75cbb5a95d217ab7b4a06494e91","internal_anchors":11},"formal_canon":{"evidence_count":2,"snapshot_sha256":"25a2d4e30e413f2993b5b0dc4bbf4653a831193a26daf8e758a4a4c926d03d5a"},"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":"2211.01324","created_at":"2026-05-17T23:39:05.095784+00:00"},{"alias_kind":"arxiv_version","alias_value":"2211.01324v5","created_at":"2026-05-17T23:39:05.095784+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.01324","created_at":"2026-05-17T23:39:05.095784+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y5HR3MRXMK6F","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y5HR3MRXMK6FSX6P","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y5HR3MRX","created_at":"2026-05-18T12:33:33.725879+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":48,"internal_anchor_count":48,"sample":[{"citing_arxiv_id":"2507.23372","citing_title":"UniEmo: Unifying Emotional Understanding and Generation with Learnable Expert Queries","ref_index":44,"is_internal_anchor":true},{"citing_arxiv_id":"2511.17171","citing_title":"FireScope: Wildfire Risk Raster Prediction with a Chain-of-Thought Oracle","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2210.00379","citing_title":"NeRF: Neural Radiance Field in 3D Vision: A Comprehensive Review (Updated Post-Gaussian Splatting)","ref_index":140,"is_internal_anchor":true},{"citing_arxiv_id":"2305.16347","citing_title":"Prompt Evolution for Generative AI: A Classifier-Guided Approach","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2304.09479","citing_title":"DiFaReli++: Diffusion Face Relighting with Consistent Cast Shadows","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2401.16764","citing_title":"BoostDream: Efficient Refining for High-Quality Text-to-3D Generation from Multi-View Diffusion","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2409.08248","citing_title":"TextBoost: Boosting Text Encoder for Personalized Text-to-Image Generation","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2603.08155","citing_title":"C$^2$FG: Control Classifier-Free Guidance via Score Discrepancy Analysis","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2604.27505","citing_title":"Leveraging Verifier-Based Reinforcement Learning in Image Editing","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2605.15684","citing_title":"ElasticDiT: Efficient Diffusion Transformers via Elastic Architecture and Sparse Attention for High-Resolution Image Generation on Mobile Devices","ref_index":50,"is_internal_anchor":true},{"citing_arxiv_id":"2308.08089","citing_title":"DragNUWA: Fine-grained Control in Video Generation by Integrating Text, Image, and Trajectory","ref_index":78,"is_internal_anchor":true},{"citing_arxiv_id":"2605.18010","citing_title":"Functionalization via Structure Completion and Motion Rectification","ref_index":235,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19532","citing_title":"Boosting Text-to-Image Diffusion Models via Core Token Attention-Based Seed Selection","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20079","citing_title":"Probability-Conserving Flow Guidance","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2605.16732","citing_title":"DiRotQ: Rotation-Aware Quantization for 4-bit Diffusion Transformers","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2507.18064","citing_title":"Adapting Large VLMs with Iterative and Manual Instructions for Generative Low-light Enhancement","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2510.21366","citing_title":"BADiff: Bandwidth Adaptive Diffusion Model","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2511.07756","citing_title":"Determinism of Randomness: Prompt-Residual Seed Shaping for Diffusion Generation","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2511.13285","citing_title":"SkyReels-Text: Fine-Grained Font-Controllable Text Editing for Poster Design","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2401.07519","citing_title":"InstantID: Zero-shot Identity-Preserving Generation in Seconds","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2511.17171","citing_title":"FireScope: Wildfire Risk Raster Prediction with a Chain-of-Thought Oracle","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2601.22160","citing_title":"Screen, Cache, and Match: A Training-Free Causality-Consistent Reference Frame Framework for Human Animation","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2302.08453","citing_title":"T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Models","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2405.08748","citing_title":"Hunyuan-DiT: A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2310.06114","citing_title":"Learning Interactive Real-World Simulators","ref_index":90,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4","json":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4.json","graph_json":"https://pith.science/api/pith-number/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/graph.json","events_json":"https://pith.science/api/pith-number/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/events.json","paper":"https://pith.science/paper/Y5HR3MRX"},"agent_actions":{"view_html":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4","download_json":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4.json","view_paper":"https://pith.science/paper/Y5HR3MRX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2211.01324&json=true","fetch_graph":"https://pith.science/api/pith-number/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/graph.json","fetch_events":"https://pith.science/api/pith-number/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/action/storage_attestation","attest_author":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/action/author_attestation","sign_citation":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/action/citation_signature","submit_replication":"https://pith.science/pith/Y5HR3MRXMK6FSX6PMAM4S5Z6Q4/action/replication_record"}},"created_at":"2026-05-17T23:39:05.095784+00:00","updated_at":"2026-05-17T23:39:05.095784+00:00"}