{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:F5R2AVOPUOZJFC3ZRCV2EW6Y7E","short_pith_number":"pith:F5R2AVOP","schema_version":"1.0","canonical_sha256":"2f63a055cfa3b2928b7988aba25bd8f921cfd764d1483ad9cd1b1ebda9dfd63d","source":{"kind":"arxiv","id":"2411.08034","version":3},"attestation_state":"computed","paper":{"title":"Scaling Properties of Diffusion Models for Perceptual Tasks","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jathushan Rajasegaran, Jitendra Malik, Rahul Ravishankar, Zeeshan Patel","submitted_at":"2024-11-12T18:59:35Z","abstract_excerpt":"In this paper, we argue that iterative computation with diffusion models offers a powerful paradigm for not only generation but also visual perception tasks. We unify tasks such as depth estimation, optical flow, and amodal segmentation under the framework of image-to-image translation, and show how diffusion models benefit from scaling training and test-time compute for these perceptual tasks. Through a careful analysis of these scaling properties, we formulate compute-optimal training and inference recipes to scale diffusion models for visual perception tasks. Our models achieve competitive "},"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":"2411.08034","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2024-11-12T18:59:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"51f19318a090155b001dbb9746fe1ac3c03811d361528beb8447a36a3eb0da07","abstract_canon_sha256":"2ae0513e9ae8e85001445012a4a307aebe557169179dc7c9174d08394cec374e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:36:16.404560Z","signature_b64":"D1lj3OT87In7vrgAAwA82kqb1V61Y+v7+t4S/0P5482XpGNs8kmcn2MPB/Pmin+l+G7B8mu9Ib7jeu9XUt0zBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2f63a055cfa3b2928b7988aba25bd8f921cfd764d1483ad9cd1b1ebda9dfd63d","last_reissued_at":"2026-07-05T09:36:16.404038Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:36:16.404038Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Scaling Properties of Diffusion Models for Perceptual Tasks","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jathushan Rajasegaran, Jitendra Malik, Rahul Ravishankar, Zeeshan Patel","submitted_at":"2024-11-12T18:59:35Z","abstract_excerpt":"In this paper, we argue that iterative computation with diffusion models offers a powerful paradigm for not only generation but also visual perception tasks. We unify tasks such as depth estimation, optical flow, and amodal segmentation under the framework of image-to-image translation, and show how diffusion models benefit from scaling training and test-time compute for these perceptual tasks. Through a careful analysis of these scaling properties, we formulate compute-optimal training and inference recipes to scale diffusion models for visual perception tasks. Our models achieve competitive "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.08034","kind":"arxiv","version":3},"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/2411.08034/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":"2411.08034","created_at":"2026-07-05T09:36:16.404102+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.08034v3","created_at":"2026-07-05T09:36:16.404102+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.08034","created_at":"2026-07-05T09:36:16.404102+00:00"},{"alias_kind":"pith_short_12","alias_value":"F5R2AVOPUOZJ","created_at":"2026-07-05T09:36:16.404102+00:00"},{"alias_kind":"pith_short_16","alias_value":"F5R2AVOPUOZJFC3Z","created_at":"2026-07-05T09:36:16.404102+00:00"},{"alias_kind":"pith_short_8","alias_value":"F5R2AVOP","created_at":"2026-07-05T09:36:16.404102+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2505.15263","citing_title":"gen2seg: Generative Models Enable Generalizable Instance Segmentation","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2507.12549","citing_title":"The Serial Scaling Hypothesis","ref_index":88,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E","json":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E.json","graph_json":"https://pith.science/api/pith-number/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/graph.json","events_json":"https://pith.science/api/pith-number/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/events.json","paper":"https://pith.science/paper/F5R2AVOP"},"agent_actions":{"view_html":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E","download_json":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E.json","view_paper":"https://pith.science/paper/F5R2AVOP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.08034&json=true","fetch_graph":"https://pith.science/api/pith-number/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/graph.json","fetch_events":"https://pith.science/api/pith-number/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/action/storage_attestation","attest_author":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/action/author_attestation","sign_citation":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/action/citation_signature","submit_replication":"https://pith.science/pith/F5R2AVOPUOZJFC3ZRCV2EW6Y7E/action/replication_record"}},"created_at":"2026-07-05T09:36:16.404102+00:00","updated_at":"2026-07-05T09:36:16.404102+00:00"}