{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ONXCDEEM7ORQGICDZHJPR6DNTF","short_pith_number":"pith:ONXCDEEM","schema_version":"1.0","canonical_sha256":"736e21908cfba3032043c9d2f8f86d997d5bf9aae79038063b60102cf2f21b17","source":{"kind":"arxiv","id":"2401.14404","version":1},"attestation_state":"computed","paper":{"title":"Deconstructing Denoising Diffusion Models for Self-Supervised Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Kaiming He, Saining Xie, Xinlei Chen, Zhuang Liu","submitted_at":"2024-01-25T18:59:57Z","abstract_excerpt":"In this study, we examine the representation learning abilities of Denoising Diffusion Models (DDM) that were originally purposed for image generation. Our philosophy is to deconstruct a DDM, gradually transforming it into a classical Denoising Autoencoder (DAE). This deconstructive procedure allows us to explore how various components of modern DDMs influence self-supervised representation learning. We observe that only a very few modern components are critical for learning good representations, while many others are nonessential. Our study ultimately arrives at an approach that is highly sim"},"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":"2401.14404","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-01-25T18:59:57Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"acfd089dc80a05d5d201516f0f68ff2f6a1caf1233424b0bfeaa5028216b2497","abstract_canon_sha256":"cb0f1a6c5e64c830fb5a9afec1377784d7a0cf5a829883496235191247c8b8b7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:37:35.163061Z","signature_b64":"Iai/KsEJoHy+8Kew5rYVpYGiX9tjiLBXDAYpMmGtC2d4KXpatRcykEsaE3wRzAoOycEKPwISlY+IAwqX7JRkAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"736e21908cfba3032043c9d2f8f86d997d5bf9aae79038063b60102cf2f21b17","last_reissued_at":"2026-07-05T07:37:35.162598Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:37:35.162598Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Deconstructing Denoising Diffusion Models for Self-Supervised Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Kaiming He, Saining Xie, Xinlei Chen, Zhuang Liu","submitted_at":"2024-01-25T18:59:57Z","abstract_excerpt":"In this study, we examine the representation learning abilities of Denoising Diffusion Models (DDM) that were originally purposed for image generation. Our philosophy is to deconstruct a DDM, gradually transforming it into a classical Denoising Autoencoder (DAE). This deconstructive procedure allows us to explore how various components of modern DDMs influence self-supervised representation learning. We observe that only a very few modern components are critical for learning good representations, while many others are nonessential. Our study ultimately arrives at an approach that is highly sim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.14404","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/2401.14404/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":"2401.14404","created_at":"2026-07-05T07:37:35.162652+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.14404v1","created_at":"2026-07-05T07:37:35.162652+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.14404","created_at":"2026-07-05T07:37:35.162652+00:00"},{"alias_kind":"pith_short_12","alias_value":"ONXCDEEM7ORQ","created_at":"2026-07-05T07:37:35.162652+00:00"},{"alias_kind":"pith_short_16","alias_value":"ONXCDEEM7ORQGICD","created_at":"2026-07-05T07:37:35.162652+00:00"},{"alias_kind":"pith_short_8","alias_value":"ONXCDEEM","created_at":"2026-07-05T07:37:35.162652+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24888","citing_title":"DiffusionBench: On Holistic Evaluation of Diffusion Transformers","ref_index":107,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18714","citing_title":"Semantic Generative Tuning for Unified Multimodal Models","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.30147","citing_title":"T2LDM++: A Self-Conditioned Representation Guided Diffusion Model for Realistic Text-to-LiDAR Scene Generation","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00583","citing_title":"Improving Visual Representation Alignment Generation with GRPO","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21502","citing_title":"VFM$^{4}$SDG: Unveiling the Power of VFMs for Single-Domain Generalized Object Detection","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18714","citing_title":"Semantic Generative Tuning for Unified Multimodal Models","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2410.06940","citing_title":"Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think","ref_index":126,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10019","citing_title":"The two clocks and the innovation window: When and how generative models learn rules","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21502","citing_title":"VFM$^{4}$SDG: Unveiling the Power of VFMs for Single-Domain Generalized Object Detection","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF","json":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF.json","graph_json":"https://pith.science/api/pith-number/ONXCDEEM7ORQGICDZHJPR6DNTF/graph.json","events_json":"https://pith.science/api/pith-number/ONXCDEEM7ORQGICDZHJPR6DNTF/events.json","paper":"https://pith.science/paper/ONXCDEEM"},"agent_actions":{"view_html":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF","download_json":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF.json","view_paper":"https://pith.science/paper/ONXCDEEM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.14404&json=true","fetch_graph":"https://pith.science/api/pith-number/ONXCDEEM7ORQGICDZHJPR6DNTF/graph.json","fetch_events":"https://pith.science/api/pith-number/ONXCDEEM7ORQGICDZHJPR6DNTF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF/action/storage_attestation","attest_author":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF/action/author_attestation","sign_citation":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF/action/citation_signature","submit_replication":"https://pith.science/pith/ONXCDEEM7ORQGICDZHJPR6DNTF/action/replication_record"}},"created_at":"2026-07-05T07:37:35.162652+00:00","updated_at":"2026-07-05T07:37:35.162652+00:00"}