{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3FFBPGSXWV7WZRVVQ3NP2FG42K","short_pith_number":"pith:3FFBPGSX","schema_version":"1.0","canonical_sha256":"d94a179a57b57f6cc6b586dafd14dcd2beed673740046d712e1f65a441e81156","source":{"kind":"arxiv","id":"2503.08253","version":1},"attestation_state":"computed","paper":{"title":"SARA: Structural and Adversarial Representation Alignment for Training-efficient Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hao Li, Hesen Chen, Junyan Wang, Zhiyu Tan","submitted_at":"2025-03-11T10:17:32Z","abstract_excerpt":"Modern diffusion models encounter a fundamental trade-off between training efficiency and generation quality. While existing representation alignment methods, such as REPA, accelerate convergence through patch-wise alignment, they often fail to capture structural relationships within visual representations and ensure global distribution consistency between pretrained encoders and denoising networks. To address these limitations, we introduce SARA, a hierarchical alignment framework that enforces multi-level representation constraints: (1) patch-wise alignment to preserve local semantic details"},"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":"2503.08253","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-11T10:17:32Z","cross_cats_sorted":[],"title_canon_sha256":"e23c62470369b7a44a7a0aa3c0b1b093fb0f2e486fd903e3f3d3bca0875e11eb","abstract_canon_sha256":"698fc0c82417de64eb084d0831a06db7713fe41cb4381504c624d7baf3dc7f21"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:28:54.765443Z","signature_b64":"s4TT4drwNuHil+jMCAJws2OGyQh4JvSxoEJujK6Rg35gWQtHZ478bqS2jFPPslwcVKaXimdDKBk5D7jbnUBbAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d94a179a57b57f6cc6b586dafd14dcd2beed673740046d712e1f65a441e81156","last_reissued_at":"2026-07-05T10:28:54.764846Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:28:54.764846Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SARA: Structural and Adversarial Representation Alignment for Training-efficient Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hao Li, Hesen Chen, Junyan Wang, Zhiyu Tan","submitted_at":"2025-03-11T10:17:32Z","abstract_excerpt":"Modern diffusion models encounter a fundamental trade-off between training efficiency and generation quality. While existing representation alignment methods, such as REPA, accelerate convergence through patch-wise alignment, they often fail to capture structural relationships within visual representations and ensure global distribution consistency between pretrained encoders and denoising networks. To address these limitations, we introduce SARA, a hierarchical alignment framework that enforces multi-level representation constraints: (1) patch-wise alignment to preserve local semantic details"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.08253","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/2503.08253/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":"2503.08253","created_at":"2026-07-05T10:28:54.764910+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.08253v1","created_at":"2026-07-05T10:28:54.764910+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.08253","created_at":"2026-07-05T10:28:54.764910+00:00"},{"alias_kind":"pith_short_12","alias_value":"3FFBPGSXWV7W","created_at":"2026-07-05T10:28:54.764910+00:00"},{"alias_kind":"pith_short_16","alias_value":"3FFBPGSXWV7WZRVV","created_at":"2026-07-05T10:28:54.764910+00:00"},{"alias_kind":"pith_short_8","alias_value":"3FFBPGSX","created_at":"2026-07-05T10:28:54.764910+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.04547","citing_title":"Stage-adaptive audio diffusion modeling","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07397","citing_title":"Data Warmup: Complexity-Aware Curricula for Efficient Diffusion Training","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K","json":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K.json","graph_json":"https://pith.science/api/pith-number/3FFBPGSXWV7WZRVVQ3NP2FG42K/graph.json","events_json":"https://pith.science/api/pith-number/3FFBPGSXWV7WZRVVQ3NP2FG42K/events.json","paper":"https://pith.science/paper/3FFBPGSX"},"agent_actions":{"view_html":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K","download_json":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K.json","view_paper":"https://pith.science/paper/3FFBPGSX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.08253&json=true","fetch_graph":"https://pith.science/api/pith-number/3FFBPGSXWV7WZRVVQ3NP2FG42K/graph.json","fetch_events":"https://pith.science/api/pith-number/3FFBPGSXWV7WZRVVQ3NP2FG42K/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K/action/storage_attestation","attest_author":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K/action/author_attestation","sign_citation":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K/action/citation_signature","submit_replication":"https://pith.science/pith/3FFBPGSXWV7WZRVVQ3NP2FG42K/action/replication_record"}},"created_at":"2026-07-05T10:28:54.764910+00:00","updated_at":"2026-07-05T10:28:54.764910+00:00"}