{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5R4YVDEOSEBGPR754SEMI26HRS","short_pith_number":"pith:5R4YVDEO","schema_version":"1.0","canonical_sha256":"ec798a8c8e910267c7fde488c46bc78c9a596fd3e1d0b6501185c9b7861b9424","source":{"kind":"arxiv","id":"2501.15061","version":2},"attestation_state":"computed","paper":{"title":"PolaFormer: Polarity-aware Linear Attention for Vision Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Dongmei Jiang, Weikang Meng, Xin Li, Yadan Luo, Zheng Zhang","submitted_at":"2025-01-25T03:46:35Z","abstract_excerpt":"Linear attention has emerged as a promising alternative to softmax-based attention, leveraging kernelized feature maps to reduce complexity from quadratic to linear in sequence length. However, the non-negative constraint on feature maps and the relaxed exponential function used in approximation lead to significant information loss compared to the original query-key dot products, resulting in less discriminative attention maps with higher entropy. To address the missing interactions driven by negative values in query-key pairs, we propose a polarity-aware linear attention mechanism that explic"},"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":"2501.15061","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-01-25T03:46:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9ce2f94796569930e34376c3749dbdec970c2eb346676ce59cf01a9a9fde006c","abstract_canon_sha256":"104a01d4aaf7cda1e3ed17e460d742664fc33024d4486c01de8a69e94c5feac3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:23:28.731929Z","signature_b64":"CqLutlVX4LGJvYsrP3D04BQa5GfxgISnaJ65vtUg0qCbWrQmq6JZSTB8v0vkaDlCtY+MGbF7l8D2DZMTBVerDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ec798a8c8e910267c7fde488c46bc78c9a596fd3e1d0b6501185c9b7861b9424","last_reissued_at":"2026-07-05T10:23:28.731243Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:23:28.731243Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PolaFormer: Polarity-aware Linear Attention for Vision Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Dongmei Jiang, Weikang Meng, Xin Li, Yadan Luo, Zheng Zhang","submitted_at":"2025-01-25T03:46:35Z","abstract_excerpt":"Linear attention has emerged as a promising alternative to softmax-based attention, leveraging kernelized feature maps to reduce complexity from quadratic to linear in sequence length. However, the non-negative constraint on feature maps and the relaxed exponential function used in approximation lead to significant information loss compared to the original query-key dot products, resulting in less discriminative attention maps with higher entropy. To address the missing interactions driven by negative values in query-key pairs, we propose a polarity-aware linear attention mechanism that explic"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.15061","kind":"arxiv","version":2},"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/2501.15061/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":"2501.15061","created_at":"2026-07-05T10:23:28.731325+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.15061v2","created_at":"2026-07-05T10:23:28.731325+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.15061","created_at":"2026-07-05T10:23:28.731325+00:00"},{"alias_kind":"pith_short_12","alias_value":"5R4YVDEOSEBG","created_at":"2026-07-05T10:23:28.731325+00:00"},{"alias_kind":"pith_short_16","alias_value":"5R4YVDEOSEBGPR75","created_at":"2026-07-05T10:23:28.731325+00:00"},{"alias_kind":"pith_short_8","alias_value":"5R4YVDEO","created_at":"2026-07-05T10:23:28.731325+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26636","citing_title":"JetViT: Efficient High-Resolution Vision Transformer with Post-Training Attention Search","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20600","citing_title":"Head-Aware Key-Value Compression for Efficient Autoregressive Image Generation","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12491","citing_title":"Elastic Attention Cores for Scalable Vision Transformers","ref_index":89,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS","json":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS.json","graph_json":"https://pith.science/api/pith-number/5R4YVDEOSEBGPR754SEMI26HRS/graph.json","events_json":"https://pith.science/api/pith-number/5R4YVDEOSEBGPR754SEMI26HRS/events.json","paper":"https://pith.science/paper/5R4YVDEO"},"agent_actions":{"view_html":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS","download_json":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS.json","view_paper":"https://pith.science/paper/5R4YVDEO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.15061&json=true","fetch_graph":"https://pith.science/api/pith-number/5R4YVDEOSEBGPR754SEMI26HRS/graph.json","fetch_events":"https://pith.science/api/pith-number/5R4YVDEOSEBGPR754SEMI26HRS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS/action/storage_attestation","attest_author":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS/action/author_attestation","sign_citation":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS/action/citation_signature","submit_replication":"https://pith.science/pith/5R4YVDEOSEBGPR754SEMI26HRS/action/replication_record"}},"created_at":"2026-07-05T10:23:28.731325+00:00","updated_at":"2026-07-05T10:23:28.731325+00:00"}