{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:V5LMAXBBRDUXZS3CQWN26C7PUD","short_pith_number":"pith:V5LMAXBB","schema_version":"1.0","canonical_sha256":"af56c05c2188e97ccb62859baf0befa0ddca4d3253547bdf25f9e7def2609577","source":{"kind":"arxiv","id":"2403.10913","version":1},"attestation_state":"computed","paper":{"title":"DEFA: Efficient Deformable Attention Acceleration via Pruning-Assisted Grid-Sampling and Multi-Scale Parallel Processing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Dongxu Lyu, Gang Wang, Guanghui He, Haomin Li, Yansong Xu, Yuzhou Chen, Zhenyu Li, Zhican Wang, Zilong Wang","submitted_at":"2024-03-16T12:34:45Z","abstract_excerpt":"Multi-scale deformable attention (MSDeformAttn) has emerged as a key mechanism in various vision tasks, demonstrating explicit superiority attributed to multi-scale grid-sampling. However, this newly introduced operator incurs irregular data access and enormous memory requirement, leading to severe PE underutilization. Meanwhile, existing approaches for attention acceleration cannot be directly applied to MSDeformAttn due to lack of support for this distinct procedure. Therefore, we propose a dedicated algorithm-architecture co-design dubbed DEFA, the first-of-its-kind method for MSDeformAttn "},"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":"2403.10913","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2024-03-16T12:34:45Z","cross_cats_sorted":[],"title_canon_sha256":"48fedef6152468dd08bb3094cfe5c8282007d84ec5bf43bf6489e3d361048fe3","abstract_canon_sha256":"e499faba32db7f1e76a90a5e0a1ae8423e80f28e6453d630814761a2f5ceeee4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:57:08.872081Z","signature_b64":"sTF1Q1QiK0VIObvR8gSfZe8jUg5/qoo4YgPIZ6+/XZjiOTF+noj5OwK5aQIUs+mM8LEgOS6iFLGe73Birp8BCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"af56c05c2188e97ccb62859baf0befa0ddca4d3253547bdf25f9e7def2609577","last_reissued_at":"2026-07-05T07:57:08.871635Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:57:08.871635Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DEFA: Efficient Deformable Attention Acceleration via Pruning-Assisted Grid-Sampling and Multi-Scale Parallel Processing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Dongxu Lyu, Gang Wang, Guanghui He, Haomin Li, Yansong Xu, Yuzhou Chen, Zhenyu Li, Zhican Wang, Zilong Wang","submitted_at":"2024-03-16T12:34:45Z","abstract_excerpt":"Multi-scale deformable attention (MSDeformAttn) has emerged as a key mechanism in various vision tasks, demonstrating explicit superiority attributed to multi-scale grid-sampling. However, this newly introduced operator incurs irregular data access and enormous memory requirement, leading to severe PE underutilization. Meanwhile, existing approaches for attention acceleration cannot be directly applied to MSDeformAttn due to lack of support for this distinct procedure. Therefore, we propose a dedicated algorithm-architecture co-design dubbed DEFA, the first-of-its-kind method for MSDeformAttn "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.10913","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/2403.10913/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":"2403.10913","created_at":"2026-07-05T07:57:08.871687+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.10913v1","created_at":"2026-07-05T07:57:08.871687+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.10913","created_at":"2026-07-05T07:57:08.871687+00:00"},{"alias_kind":"pith_short_12","alias_value":"V5LMAXBBRDUX","created_at":"2026-07-05T07:57:08.871687+00:00"},{"alias_kind":"pith_short_16","alias_value":"V5LMAXBBRDUXZS3C","created_at":"2026-07-05T07:57:08.871687+00:00"},{"alias_kind":"pith_short_8","alias_value":"V5LMAXBB","created_at":"2026-07-05T07:57:08.871687+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD","json":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD.json","graph_json":"https://pith.science/api/pith-number/V5LMAXBBRDUXZS3CQWN26C7PUD/graph.json","events_json":"https://pith.science/api/pith-number/V5LMAXBBRDUXZS3CQWN26C7PUD/events.json","paper":"https://pith.science/paper/V5LMAXBB"},"agent_actions":{"view_html":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD","download_json":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD.json","view_paper":"https://pith.science/paper/V5LMAXBB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.10913&json=true","fetch_graph":"https://pith.science/api/pith-number/V5LMAXBBRDUXZS3CQWN26C7PUD/graph.json","fetch_events":"https://pith.science/api/pith-number/V5LMAXBBRDUXZS3CQWN26C7PUD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD/action/storage_attestation","attest_author":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD/action/author_attestation","sign_citation":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD/action/citation_signature","submit_replication":"https://pith.science/pith/V5LMAXBBRDUXZS3CQWN26C7PUD/action/replication_record"}},"created_at":"2026-07-05T07:57:08.871687+00:00","updated_at":"2026-07-05T07:57:08.871687+00:00"}