{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:VVBAPBK5DFG6CRIWZ3O6XEDJRM","short_pith_number":"pith:VVBAPBK5","schema_version":"1.0","canonical_sha256":"ad4207855d194de14516ceddeb90698b3b33f129a2bb5335175e43185f49c136","source":{"kind":"arxiv","id":"2001.01037","version":5},"attestation_state":"computed","paper":{"title":"Explain and Improve: LRP-Inference Fine-Tuning for Image Captioning Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexander Binder, Jiamei Sun, Sebastian Lapuschkin, Wojciech Samek","submitted_at":"2020-01-04T05:15:11Z","abstract_excerpt":"This paper analyzes the predictions of image captioning models with attention mechanisms beyond visualizing the attention itself. We develop variants of layer-wise relevance propagation (LRP) and gradient-based explanation methods, tailored to image captioning models with attention mechanisms. We compare the interpretability of attention heatmaps systematically against the explanations provided by explanation methods such as LRP, Grad-CAM, and Guided Grad-CAM. We show that explanation methods provide simultaneously pixel-wise image explanations (supporting and opposing pixels of the input imag"},"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":"2001.01037","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2020-01-04T05:15:11Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"75c6656ceb15a61131a60028e39159db5dce37388975ff004216880f900dfbcf","abstract_canon_sha256":"ef345bc57ca7c1c252883f97241ec907935a8f54db49cf49223204f1cf21191d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:02:07.019753Z","signature_b64":"62u/1TT6BMbZ0paZqEhV8LHGYhMbZmrOJqqYb6dVoFHk+mKB1WXtZQeJ+bsUAzMSGfKFQT3OpKd/5vJ8hviuBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ad4207855d194de14516ceddeb90698b3b33f129a2bb5335175e43185f49c136","last_reissued_at":"2026-07-05T03:02:07.019343Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:02:07.019343Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Explain and Improve: LRP-Inference Fine-Tuning for Image Captioning Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexander Binder, Jiamei Sun, Sebastian Lapuschkin, Wojciech Samek","submitted_at":"2020-01-04T05:15:11Z","abstract_excerpt":"This paper analyzes the predictions of image captioning models with attention mechanisms beyond visualizing the attention itself. We develop variants of layer-wise relevance propagation (LRP) and gradient-based explanation methods, tailored to image captioning models with attention mechanisms. We compare the interpretability of attention heatmaps systematically against the explanations provided by explanation methods such as LRP, Grad-CAM, and Guided Grad-CAM. We show that explanation methods provide simultaneously pixel-wise image explanations (supporting and opposing pixels of the input imag"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2001.01037","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/2001.01037/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":"2001.01037","created_at":"2026-07-05T03:02:07.019405+00:00"},{"alias_kind":"arxiv_version","alias_value":"2001.01037v5","created_at":"2026-07-05T03:02:07.019405+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2001.01037","created_at":"2026-07-05T03:02:07.019405+00:00"},{"alias_kind":"pith_short_12","alias_value":"VVBAPBK5DFG6","created_at":"2026-07-05T03:02:07.019405+00:00"},{"alias_kind":"pith_short_16","alias_value":"VVBAPBK5DFG6CRIW","created_at":"2026-07-05T03:02:07.019405+00:00"},{"alias_kind":"pith_short_8","alias_value":"VVBAPBK5","created_at":"2026-07-05T03:02:07.019405+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.21246","citing_title":"On Explaining Visual Captioning with Hybrid Markov Logic Networks","ref_index":31,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM","json":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM.json","graph_json":"https://pith.science/api/pith-number/VVBAPBK5DFG6CRIWZ3O6XEDJRM/graph.json","events_json":"https://pith.science/api/pith-number/VVBAPBK5DFG6CRIWZ3O6XEDJRM/events.json","paper":"https://pith.science/paper/VVBAPBK5"},"agent_actions":{"view_html":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM","download_json":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM.json","view_paper":"https://pith.science/paper/VVBAPBK5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2001.01037&json=true","fetch_graph":"https://pith.science/api/pith-number/VVBAPBK5DFG6CRIWZ3O6XEDJRM/graph.json","fetch_events":"https://pith.science/api/pith-number/VVBAPBK5DFG6CRIWZ3O6XEDJRM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM/action/storage_attestation","attest_author":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM/action/author_attestation","sign_citation":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM/action/citation_signature","submit_replication":"https://pith.science/pith/VVBAPBK5DFG6CRIWZ3O6XEDJRM/action/replication_record"}},"created_at":"2026-07-05T03:02:07.019405+00:00","updated_at":"2026-07-05T03:02:07.019405+00:00"}