{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:CEVIJTLKVV2YJZ7BZBJ2DZKJFL","short_pith_number":"pith:CEVIJTLK","schema_version":"1.0","canonical_sha256":"112a84cd6aad7584e7e1c853a1e5492ae959aadcdd74d15f21407e897450efe9","source":{"kind":"arxiv","id":"2103.02152","version":3},"attestation_state":"computed","paper":{"title":"Group-wise Inhibition based Feature Regularization for Robust Classification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Feng Liu, Haoqian Wu, Haozhe Liu, Linlin Shen, Weicheng Xie","submitted_at":"2021-03-03T03:19:32Z","abstract_excerpt":"The convolutional neural network (CNN) is vulnerable to degraded images with even very small variations (e.g. corrupted and adversarial samples). One of the possible reasons is that CNN pays more attention to the most discriminative regions, but ignores the auxiliary features when learning, leading to the lack of feature diversity for final judgment. In our method, we propose to dynamically suppress significant activation values of CNN by group-wise inhibition, but not fixedly or randomly handle them when training. The feature maps with different activation distribution are then processed sepa"},"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":"2103.02152","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-03-03T03:19:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4aacb5211495f222ca0fe590483707711813d353506b84c3adfb1e9fb6aa80a6","abstract_canon_sha256":"2db3a946bd03315dea804dc3413cd79ecb4208a5d80c2137a53e33d07be30d88"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:06:20.910590Z","signature_b64":"09bJ/hEB7BTrI5GQe0l3du5u4lBw8G78x1XIZ+0rKfnK6H0aR+cRxfSNINbXPgOOVX5ffNE4XYJAn8AS9x9pAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"112a84cd6aad7584e7e1c853a1e5492ae959aadcdd74d15f21407e897450efe9","last_reissued_at":"2026-07-05T03:06:20.910165Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:06:20.910165Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Group-wise Inhibition based Feature Regularization for Robust Classification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Feng Liu, Haoqian Wu, Haozhe Liu, Linlin Shen, Weicheng Xie","submitted_at":"2021-03-03T03:19:32Z","abstract_excerpt":"The convolutional neural network (CNN) is vulnerable to degraded images with even very small variations (e.g. corrupted and adversarial samples). One of the possible reasons is that CNN pays more attention to the most discriminative regions, but ignores the auxiliary features when learning, leading to the lack of feature diversity for final judgment. In our method, we propose to dynamically suppress significant activation values of CNN by group-wise inhibition, but not fixedly or randomly handle them when training. The feature maps with different activation distribution are then processed sepa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.02152","kind":"arxiv","version":3},"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/2103.02152/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":"2103.02152","created_at":"2026-07-05T03:06:20.910221+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.02152v3","created_at":"2026-07-05T03:06:20.910221+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.02152","created_at":"2026-07-05T03:06:20.910221+00:00"},{"alias_kind":"pith_short_12","alias_value":"CEVIJTLKVV2Y","created_at":"2026-07-05T03:06:20.910221+00:00"},{"alias_kind":"pith_short_16","alias_value":"CEVIJTLKVV2YJZ7B","created_at":"2026-07-05T03:06:20.910221+00:00"},{"alias_kind":"pith_short_8","alias_value":"CEVIJTLK","created_at":"2026-07-05T03:06:20.910221+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/CEVIJTLKVV2YJZ7BZBJ2DZKJFL","json":"https://pith.science/pith/CEVIJTLKVV2YJZ7BZBJ2DZKJFL.json","graph_json":"https://pith.science/api/pith-number/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/graph.json","events_json":"https://pith.science/api/pith-number/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/events.json","paper":"https://pith.science/paper/CEVIJTLK"},"agent_actions":{"view_html":"https://pith.science/pith/CEVIJTLKVV2YJZ7BZBJ2DZKJFL","download_json":"https://pith.science/pith/CEVIJTLKVV2YJZ7BZBJ2DZKJFL.json","view_paper":"https://pith.science/paper/CEVIJTLK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.02152&json=true","fetch_graph":"https://pith.science/api/pith-number/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/graph.json","fetch_events":"https://pith.science/api/pith-number/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/action/storage_attestation","attest_author":"https://pith.science/pith/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/action/author_attestation","sign_citation":"https://pith.science/pith/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/action/citation_signature","submit_replication":"https://pith.science/pith/CEVIJTLKVV2YJZ7BZBJ2DZKJFL/action/replication_record"}},"created_at":"2026-07-05T03:06:20.910221+00:00","updated_at":"2026-07-05T03:06:20.910221+00:00"}