{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2018:4ZP7HHMEW4LEL7ZTDVKYJLRLAY","short_pith_number":"pith:4ZP7HHME","canonical_record":{"source":{"id":"1803.03880","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-03-11T02:02:46Z","cross_cats_sorted":["cs.IT","cs.LG","math.IT"],"title_canon_sha256":"c81e039be47666c6051382a033240e07728081fdd76f7e51a2e1b4748299d632","abstract_canon_sha256":"5a0002a3d2e0f39fa05ab341e58e85246cc18f94d607fcaca24a47cd29aa3d92"},"schema_version":"1.0"},"canonical_sha256":"e65ff39d84b71645ff331d5584ae2b062f3a0b8ee67bd95d671d10deebea492d","source":{"kind":"arxiv","id":"1803.03880","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1803.03880","created_at":"2026-05-18T00:10:48Z"},{"alias_kind":"arxiv_version","alias_value":"1803.03880v3","created_at":"2026-05-18T00:10:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1803.03880","created_at":"2026-05-18T00:10:48Z"},{"alias_kind":"pith_short_12","alias_value":"4ZP7HHMEW4LE","created_at":"2026-05-18T12:32:05Z"},{"alias_kind":"pith_short_16","alias_value":"4ZP7HHMEW4LEL7ZT","created_at":"2026-05-18T12:32:05Z"},{"alias_kind":"pith_short_8","alias_value":"4ZP7HHME","created_at":"2026-05-18T12:32:05Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2018:4ZP7HHMEW4LEL7ZTDVKYJLRLAY","target":"record","payload":{"canonical_record":{"source":{"id":"1803.03880","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-03-11T02:02:46Z","cross_cats_sorted":["cs.IT","cs.LG","math.IT"],"title_canon_sha256":"c81e039be47666c6051382a033240e07728081fdd76f7e51a2e1b4748299d632","abstract_canon_sha256":"5a0002a3d2e0f39fa05ab341e58e85246cc18f94d607fcaca24a47cd29aa3d92"},"schema_version":"1.0"},"canonical_sha256":"e65ff39d84b71645ff331d5584ae2b062f3a0b8ee67bd95d671d10deebea492d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:10:48.812687Z","signature_b64":"h0ZkWlmJFwvCZjNrRADbwcjejKcrGr8g3wED2xTmbidoX2ziUoZPRbLeNaYuxsHeoYxqoGmudMcCCS1dsL4tBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e65ff39d84b71645ff331d5584ae2b062f3a0b8ee67bd95d671d10deebea492d","last_reissued_at":"2026-05-18T00:10:48.812044Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:10:48.812044Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1803.03880","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:10:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mJaytAj/LYAUJC/M1EZwcRt5f1AmtICa2pORmzCU0tz0mNNJ48gk0VFLrCuRP9gm+EkTQtjuDglwvXE09RxBBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T15:45:15.152510Z"},"content_sha256":"a7627a8a4c657dc0a1b1437a60f6505e498c7db4c419661303f8784cfde34517","schema_version":"1.0","event_id":"sha256:a7627a8a4c657dc0a1b1437a60f6505e498c7db4c419661303f8784cfde34517"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2018:4ZP7HHMEW4LEL7ZTDVKYJLRLAY","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Combating Adversarial Attacks Using Sparse Representations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IT","cs.LG","math.IT"],"primary_cat":"stat.ML","authors_text":"Ramtin Pedarsani, Soorya Gopalakrishnan, Upamanyu Madhow, Zhinus Marzi","submitted_at":"2018-03-11T02:02:46Z","abstract_excerpt":"It is by now well-known that small adversarial perturbations can induce classification errors in deep neural networks (DNNs). In this paper, we make the case that sparse representations of the input data are a crucial tool for combating such attacks. For linear classifiers, we show that a sparsifying front end is provably effective against $\\ell_{\\infty}$-bounded attacks, reducing output distortion due to the attack by a factor of roughly $K / N$ where $N$ is the data dimension and $K$ is the sparsity level. We then extend this concept to DNNs, showing that a \"locally linear\" model can be used"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1803.03880","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":""},"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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:10:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JgcGjNtz8KxwmM2HNwCA5n/f6YhxJZ/ZmHhpPNpTGg7puxOLItyR0kTmOYAhyBWOIq6+jZZ0Cw0H0LC/hOMtBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T15:45:15.153138Z"},"content_sha256":"e2b0a306c1ea5b0e7d4a36b896ed32d2cb14fa1384097118f49760f3f3258aed","schema_version":"1.0","event_id":"sha256:e2b0a306c1ea5b0e7d4a36b896ed32d2cb14fa1384097118f49760f3f3258aed"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4ZP7HHMEW4LEL7ZTDVKYJLRLAY/bundle.json","state_url":"https://pith.science/pith/4ZP7HHMEW4LEL7ZTDVKYJLRLAY/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4ZP7HHMEW4LEL7ZTDVKYJLRLAY/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-21T15:45:15Z","links":{"resolver":"https://pith.science/pith/4ZP7HHMEW4LEL7ZTDVKYJLRLAY","bundle":"https://pith.science/pith/4ZP7HHMEW4LEL7ZTDVKYJLRLAY/bundle.json","state":"https://pith.science/pith/4ZP7HHMEW4LEL7ZTDVKYJLRLAY/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4ZP7HHMEW4LEL7ZTDVKYJLRLAY/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2018:4ZP7HHMEW4LEL7ZTDVKYJLRLAY","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"5a0002a3d2e0f39fa05ab341e58e85246cc18f94d607fcaca24a47cd29aa3d92","cross_cats_sorted":["cs.IT","cs.LG","math.IT"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-03-11T02:02:46Z","title_canon_sha256":"c81e039be47666c6051382a033240e07728081fdd76f7e51a2e1b4748299d632"},"schema_version":"1.0","source":{"id":"1803.03880","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1803.03880","created_at":"2026-05-18T00:10:48Z"},{"alias_kind":"arxiv_version","alias_value":"1803.03880v3","created_at":"2026-05-18T00:10:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1803.03880","created_at":"2026-05-18T00:10:48Z"},{"alias_kind":"pith_short_12","alias_value":"4ZP7HHMEW4LE","created_at":"2026-05-18T12:32:05Z"},{"alias_kind":"pith_short_16","alias_value":"4ZP7HHMEW4LEL7ZT","created_at":"2026-05-18T12:32:05Z"},{"alias_kind":"pith_short_8","alias_value":"4ZP7HHME","created_at":"2026-05-18T12:32:05Z"}],"graph_snapshots":[{"event_id":"sha256:e2b0a306c1ea5b0e7d4a36b896ed32d2cb14fa1384097118f49760f3f3258aed","target":"graph","created_at":"2026-05-18T00:10:48Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"It is by now well-known that small adversarial perturbations can induce classification errors in deep neural networks (DNNs). In this paper, we make the case that sparse representations of the input data are a crucial tool for combating such attacks. For linear classifiers, we show that a sparsifying front end is provably effective against $\\ell_{\\infty}$-bounded attacks, reducing output distortion due to the attack by a factor of roughly $K / N$ where $N$ is the data dimension and $K$ is the sparsity level. We then extend this concept to DNNs, showing that a \"locally linear\" model can be used","authors_text":"Ramtin Pedarsani, Soorya Gopalakrishnan, Upamanyu Madhow, Zhinus Marzi","cross_cats":["cs.IT","cs.LG","math.IT"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-03-11T02:02:46Z","title":"Combating Adversarial Attacks Using Sparse Representations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1803.03880","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:a7627a8a4c657dc0a1b1437a60f6505e498c7db4c419661303f8784cfde34517","target":"record","created_at":"2026-05-18T00:10:48Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"5a0002a3d2e0f39fa05ab341e58e85246cc18f94d607fcaca24a47cd29aa3d92","cross_cats_sorted":["cs.IT","cs.LG","math.IT"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-03-11T02:02:46Z","title_canon_sha256":"c81e039be47666c6051382a033240e07728081fdd76f7e51a2e1b4748299d632"},"schema_version":"1.0","source":{"id":"1803.03880","kind":"arxiv","version":3}},"canonical_sha256":"e65ff39d84b71645ff331d5584ae2b062f3a0b8ee67bd95d671d10deebea492d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e65ff39d84b71645ff331d5584ae2b062f3a0b8ee67bd95d671d10deebea492d","first_computed_at":"2026-05-18T00:10:48.812044Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T00:10:48.812044Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"h0ZkWlmJFwvCZjNrRADbwcjejKcrGr8g3wED2xTmbidoX2ziUoZPRbLeNaYuxsHeoYxqoGmudMcCCS1dsL4tBw==","signature_status":"signed_v1","signed_at":"2026-05-18T00:10:48.812687Z","signed_message":"canonical_sha256_bytes"},"source_id":"1803.03880","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a7627a8a4c657dc0a1b1437a60f6505e498c7db4c419661303f8784cfde34517","sha256:e2b0a306c1ea5b0e7d4a36b896ed32d2cb14fa1384097118f49760f3f3258aed"],"state_sha256":"6f84cc637115f924142acc70e7d66409877f56c0df46277f7d4df5e5e760808a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lUdgcPV4Av/iNIxpaRM8RxA1R63bLVbM91xtbxGEVW/aIanYbt+9hMKNRYa646sdDPi3qfxTj35aOCd4+iONBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-21T15:45:15.156178Z","bundle_sha256":"6aa53f29d14de20c92e35b0adfff46843bd548b64c5b441ed383a7ff1340e4c7"}}