{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:MFU6MCSSDYRALJUKJUXMPWY73D","short_pith_number":"pith:MFU6MCSS","schema_version":"1.0","canonical_sha256":"6169e60a521e2205a68a4d2ec7db1fd8f7597c6a636b28856d3259689bbd194c","source":{"kind":"arxiv","id":"1610.02391","version":4},"attestation_state":"computed","paper":{"title":"Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Abhishek Das, Devi Parikh, Dhruv Batra, Michael Cogswell, Ramakrishna Vedantam, Ramprasaath R. Selvaraju","submitted_at":"2016-10-07T19:54:24Z","abstract_excerpt":"We propose a technique for producing \"visual explanations\" for decisions from a large class of CNN-based models, making them more transparent. Our approach - Gradient-weighted Class Activation Mapping (Grad-CAM), uses the gradients of any target concept, flowing into the final convolutional layer to produce a coarse localization map highlighting important regions in the image for predicting the concept. Grad-CAM is applicable to a wide variety of CNN model-families: (1) CNNs with fully-connected layers, (2) CNNs used for structured outputs, (3) CNNs used in tasks with multimodal inputs or rein"},"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":"1610.02391","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2016-10-07T19:54:24Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"23168bad9f48e91b6fa608a136cf8ecc184e978d73d9f1c24f2cb8e4605da43c","abstract_canon_sha256":"81ec1ef99b4a4715d3a2432e0ba78abcb7026eb4ce681a8d677fe5f73655e462"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:23:19.137180Z","signature_b64":"wC6462lDm6gAbLXqaYr6Fr46zHtfXO3/JAX5jF11Pr9u7MskwXH9M1dvy4kQpGiEQIn1Hxw0Mz87C4Lu14hsAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6169e60a521e2205a68a4d2ec7db1fd8f7597c6a636b28856d3259689bbd194c","last_reissued_at":"2026-07-05T00:23:19.136615Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:23:19.136615Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Abhishek Das, Devi Parikh, Dhruv Batra, Michael Cogswell, Ramakrishna Vedantam, Ramprasaath R. Selvaraju","submitted_at":"2016-10-07T19:54:24Z","abstract_excerpt":"We propose a technique for producing \"visual explanations\" for decisions from a large class of CNN-based models, making them more transparent. Our approach - Gradient-weighted Class Activation Mapping (Grad-CAM), uses the gradients of any target concept, flowing into the final convolutional layer to produce a coarse localization map highlighting important regions in the image for predicting the concept. Grad-CAM is applicable to a wide variety of CNN model-families: (1) CNNs with fully-connected layers, (2) CNNs used for structured outputs, (3) CNNs used in tasks with multimodal inputs or rein"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1610.02391","kind":"arxiv","version":4},"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/1610.02391/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":"1610.02391","created_at":"2026-07-05T00:23:19.136687+00:00"},{"alias_kind":"arxiv_version","alias_value":"1610.02391v4","created_at":"2026-07-05T00:23:19.136687+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1610.02391","created_at":"2026-07-05T00:23:19.136687+00:00"},{"alias_kind":"pith_short_12","alias_value":"MFU6MCSSDYRA","created_at":"2026-07-05T00:23:19.136687+00:00"},{"alias_kind":"pith_short_16","alias_value":"MFU6MCSSDYRALJUK","created_at":"2026-07-05T00:23:19.136687+00:00"},{"alias_kind":"pith_short_8","alias_value":"MFU6MCSS","created_at":"2026-07-05T00:23:19.136687+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":11,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22858","citing_title":"The Unseen Hand: Manipulating Model Fairness and SHAP with Targeted Identity Re-Association Attacks","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22447","citing_title":"A Differentiable Atari VCS:A Complex, Fully Known Ground Truth for Explainable AI","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08032","citing_title":"Variational Proximal Policy Optimization","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07180","citing_title":"OPTIMUS-Prime: Minimal and Sufficient Concept Explanations for Deep Vision Models","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00385","citing_title":"MalariAI: A Label-Resilient Decoupled Framework for Annotation-Agnostic Cell Segmentation and Explainable Stage Classification in Dense Malaria Blood Smears","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"1907.02519","citing_title":"Neuron ranking -- an informed way to condense convolutional neural networks architecture","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2505.19583","citing_title":"Identifying lopsidedness in spiral galaxies using a Deep Convolutional Neural Network","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2303.08128","citing_title":"ViperGPT: Visual Inference via Python Execution for Reasoning","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2512.07420","citing_title":"KIGNet: Physics-Motivated Multi-Graph Representation Learning for Explainable Jet Tagging","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2309.16588","citing_title":"Vision Transformers Need Registers","ref_index":219,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04904","citing_title":"Exploring Clustering Capability of Inpainting Model Embeddings for Pattern-based Individual Identification","ref_index":67,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D","json":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D.json","graph_json":"https://pith.science/api/pith-number/MFU6MCSSDYRALJUKJUXMPWY73D/graph.json","events_json":"https://pith.science/api/pith-number/MFU6MCSSDYRALJUKJUXMPWY73D/events.json","paper":"https://pith.science/paper/MFU6MCSS"},"agent_actions":{"view_html":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D","download_json":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D.json","view_paper":"https://pith.science/paper/MFU6MCSS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1610.02391&json=true","fetch_graph":"https://pith.science/api/pith-number/MFU6MCSSDYRALJUKJUXMPWY73D/graph.json","fetch_events":"https://pith.science/api/pith-number/MFU6MCSSDYRALJUKJUXMPWY73D/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D/action/storage_attestation","attest_author":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D/action/author_attestation","sign_citation":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D/action/citation_signature","submit_replication":"https://pith.science/pith/MFU6MCSSDYRALJUKJUXMPWY73D/action/replication_record"}},"created_at":"2026-07-05T00:23:19.136687+00:00","updated_at":"2026-07-05T00:23:19.136687+00:00"}