{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:CJY2JZTOP3EODFFBCPKVXQVUBI","short_pith_number":"pith:CJY2JZTO","schema_version":"1.0","canonical_sha256":"1271a4e66e7ec8e194a113d55bc2b40a0b941565fa9a17410c8f654d4e16114f","source":{"kind":"arxiv","id":"2403.15603","version":2},"attestation_state":"computed","paper":{"title":"Forward Learning for Gradient-based Black-box Saliency Map Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chenliang Xu, Jinyang Jiang, Mingqian Feng, Rongyi Zhu, Yijie Peng, Zeliang Zhang","submitted_at":"2024-03-22T20:11:19Z","abstract_excerpt":"Gradient-based saliency maps are widely used to explain deep neural network decisions. However, as models become deeper and more black-box, such as in closed-source APIs like ChatGPT, computing gradients become challenging, hindering conventional explanation methods. In this work, we introduce a novel unified framework for estimating gradients in black-box settings and generating saliency maps to interpret model decisions. We employ the likelihood ratio method to estimate output-to-input gradients and utilize them for saliency map generation. Additionally, we propose blockwise computation tech"},"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":"2403.15603","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-03-22T20:11:19Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a1a3090123130d74bed852587663b537d811fbd1f5c86794104093a1d052bc2a","abstract_canon_sha256":"2ce6214110b4cc4c6d38997d512d62bb40db125576fc32c7feb60574e500e8ea"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:39:20.344986Z","signature_b64":"zcqbj6Ygp35a1EYOWdDq4BOU6ngV/yRj13FCIY5fcO6EoHXu9QTnF6aeu/W+/U4+Msci+BcTRFzuMZxa180gBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1271a4e66e7ec8e194a113d55bc2b40a0b941565fa9a17410c8f654d4e16114f","last_reissued_at":"2026-07-05T08:39:20.344607Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:39:20.344607Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Forward Learning for Gradient-based Black-box Saliency Map Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chenliang Xu, Jinyang Jiang, Mingqian Feng, Rongyi Zhu, Yijie Peng, Zeliang Zhang","submitted_at":"2024-03-22T20:11:19Z","abstract_excerpt":"Gradient-based saliency maps are widely used to explain deep neural network decisions. However, as models become deeper and more black-box, such as in closed-source APIs like ChatGPT, computing gradients become challenging, hindering conventional explanation methods. In this work, we introduce a novel unified framework for estimating gradients in black-box settings and generating saliency maps to interpret model decisions. We employ the likelihood ratio method to estimate output-to-input gradients and utilize them for saliency map generation. Additionally, we propose blockwise computation tech"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.15603","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/2403.15603/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":"2403.15603","created_at":"2026-07-05T08:39:20.344662+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.15603v2","created_at":"2026-07-05T08:39:20.344662+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.15603","created_at":"2026-07-05T08:39:20.344662+00:00"},{"alias_kind":"pith_short_12","alias_value":"CJY2JZTOP3EO","created_at":"2026-07-05T08:39:20.344662+00:00"},{"alias_kind":"pith_short_16","alias_value":"CJY2JZTOP3EODFFB","created_at":"2026-07-05T08:39:20.344662+00:00"},{"alias_kind":"pith_short_8","alias_value":"CJY2JZTO","created_at":"2026-07-05T08:39:20.344662+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/CJY2JZTOP3EODFFBCPKVXQVUBI","json":"https://pith.science/pith/CJY2JZTOP3EODFFBCPKVXQVUBI.json","graph_json":"https://pith.science/api/pith-number/CJY2JZTOP3EODFFBCPKVXQVUBI/graph.json","events_json":"https://pith.science/api/pith-number/CJY2JZTOP3EODFFBCPKVXQVUBI/events.json","paper":"https://pith.science/paper/CJY2JZTO"},"agent_actions":{"view_html":"https://pith.science/pith/CJY2JZTOP3EODFFBCPKVXQVUBI","download_json":"https://pith.science/pith/CJY2JZTOP3EODFFBCPKVXQVUBI.json","view_paper":"https://pith.science/paper/CJY2JZTO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.15603&json=true","fetch_graph":"https://pith.science/api/pith-number/CJY2JZTOP3EODFFBCPKVXQVUBI/graph.json","fetch_events":"https://pith.science/api/pith-number/CJY2JZTOP3EODFFBCPKVXQVUBI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CJY2JZTOP3EODFFBCPKVXQVUBI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CJY2JZTOP3EODFFBCPKVXQVUBI/action/storage_attestation","attest_author":"https://pith.science/pith/CJY2JZTOP3EODFFBCPKVXQVUBI/action/author_attestation","sign_citation":"https://pith.science/pith/CJY2JZTOP3EODFFBCPKVXQVUBI/action/citation_signature","submit_replication":"https://pith.science/pith/CJY2JZTOP3EODFFBCPKVXQVUBI/action/replication_record"}},"created_at":"2026-07-05T08:39:20.344662+00:00","updated_at":"2026-07-05T08:39:20.344662+00:00"}