{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3NU5ELTRPKLCC5PIKHITL75CLL","short_pith_number":"pith:3NU5ELTR","schema_version":"1.0","canonical_sha256":"db69d22e717a962175e851d135ffa25ac3aebfca2d83961eb17bb5f5ecacf57c","source":{"kind":"arxiv","id":"2302.03985","version":5},"attestation_state":"computed","paper":{"title":"Cross-Layer Retrospective Retrieving via Layer Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guangjian Tian, Guodong Li, Jingyu Zhao, Jintai Chen, Yanwen Fang, Yuxi Cai","submitted_at":"2023-02-08T10:50:01Z","abstract_excerpt":"More and more evidence has shown that strengthening layer interactions can enhance the representation power of a deep neural network, while self-attention excels at learning interdependencies by retrieving query-activated information. Motivated by this, we devise a cross-layer attention mechanism, called multi-head recurrent layer attention (MRLA), that sends a query representation of the current layer to all previous layers to retrieve query-related information from different levels of receptive fields. A light-weighted version of MRLA is also proposed to reduce the quadratic computation cost"},"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":"2302.03985","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-02-08T10:50:01Z","cross_cats_sorted":[],"title_canon_sha256":"6815a638f50430a259d79bbac0df0d549d0a9816aab4034f997fd62c563a67a6","abstract_canon_sha256":"42755ec52b3d55abae2069d31d02eae6b7a1fad7f2dffadf100ce82859b12879"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:46:37.461960Z","signature_b64":"r76Dmt+BmWF307qVs7ZnZOgOaCoKHDeczryll/IN3vg0xPeYp6rX7K4XngNem1D5ErCRPwMXQiE+xm3Q+7fhCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"db69d22e717a962175e851d135ffa25ac3aebfca2d83961eb17bb5f5ecacf57c","last_reissued_at":"2026-07-05T05:46:37.461467Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:46:37.461467Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Cross-Layer Retrospective Retrieving via Layer Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guangjian Tian, Guodong Li, Jingyu Zhao, Jintai Chen, Yanwen Fang, Yuxi Cai","submitted_at":"2023-02-08T10:50:01Z","abstract_excerpt":"More and more evidence has shown that strengthening layer interactions can enhance the representation power of a deep neural network, while self-attention excels at learning interdependencies by retrieving query-activated information. Motivated by this, we devise a cross-layer attention mechanism, called multi-head recurrent layer attention (MRLA), that sends a query representation of the current layer to all previous layers to retrieve query-related information from different levels of receptive fields. A light-weighted version of MRLA is also proposed to reduce the quadratic computation cost"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.03985","kind":"arxiv","version":5},"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/2302.03985/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":"2302.03985","created_at":"2026-07-05T05:46:37.461526+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.03985v5","created_at":"2026-07-05T05:46:37.461526+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.03985","created_at":"2026-07-05T05:46:37.461526+00:00"},{"alias_kind":"pith_short_12","alias_value":"3NU5ELTRPKLC","created_at":"2026-07-05T05:46:37.461526+00:00"},{"alias_kind":"pith_short_16","alias_value":"3NU5ELTRPKLCC5PI","created_at":"2026-07-05T05:46:37.461526+00:00"},{"alias_kind":"pith_short_8","alias_value":"3NU5ELTR","created_at":"2026-07-05T05:46:37.461526+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.00057","citing_title":"Enhancing Oracle Bone Inscription Recognition via Multi-Scale Layer Attention","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2603.15031","citing_title":"Attention Residuals","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03953","citing_title":"Transformers with Selective Access to Early Representations","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03953","citing_title":"Transformers with Selective Access to Early Representations","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL","json":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL.json","graph_json":"https://pith.science/api/pith-number/3NU5ELTRPKLCC5PIKHITL75CLL/graph.json","events_json":"https://pith.science/api/pith-number/3NU5ELTRPKLCC5PIKHITL75CLL/events.json","paper":"https://pith.science/paper/3NU5ELTR"},"agent_actions":{"view_html":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL","download_json":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL.json","view_paper":"https://pith.science/paper/3NU5ELTR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.03985&json=true","fetch_graph":"https://pith.science/api/pith-number/3NU5ELTRPKLCC5PIKHITL75CLL/graph.json","fetch_events":"https://pith.science/api/pith-number/3NU5ELTRPKLCC5PIKHITL75CLL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL/action/storage_attestation","attest_author":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL/action/author_attestation","sign_citation":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL/action/citation_signature","submit_replication":"https://pith.science/pith/3NU5ELTRPKLCC5PIKHITL75CLL/action/replication_record"}},"created_at":"2026-07-05T05:46:37.461526+00:00","updated_at":"2026-07-05T05:46:37.461526+00:00"}