{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:6QAPZG3WJDHAIZJUQHDVA2WQYT","short_pith_number":"pith:6QAPZG3W","schema_version":"1.0","canonical_sha256":"f400fc9b7648ce04653481c7506ad0c4c9c7a5f1d2be4210cfcf5dbc5cd9e5e9","source":{"kind":"arxiv","id":"2003.04985","version":1},"attestation_state":"computed","paper":{"title":"Adv-BERT: BERT is not robust on misspellings! Generating nature adversarial samples on BERT","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Akari Asai, Caiming Xiong, Jia Li, Kazuma Hashimoto, Lichao Sun, Philip Yu, Wenpeng Yin","submitted_at":"2020-02-27T22:07:11Z","abstract_excerpt":"There is an increasing amount of literature that claims the brittleness of deep neural networks in dealing with adversarial examples that are created maliciously. It is unclear, however, how the models will perform in realistic scenarios where \\textit{natural rather than malicious} adversarial instances often exist. This work systematically explores the robustness of BERT, the state-of-the-art Transformer-style model in NLP, in dealing with noisy data, particularly mistakes in typing the keyboard, that occur inadvertently. Intensive experiments on sentiment analysis and question answering benc"},"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":"2003.04985","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-27T22:07:11Z","cross_cats_sorted":[],"title_canon_sha256":"5583b8bc71ef8546708b8813891cead683879aa4fc231e6cf0d22f29894f83a0","abstract_canon_sha256":"dc881447db8f4f15871a8d30164c073e7e3682e1f3dc8059aa33bbce572f2b4e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:47:20.014538Z","signature_b64":"wwxAwTk2308ZwNh4n+aEHciwgz5JNLvCTLg0J0HgifDfoCe/GjqqZuCtIUtjAzZa/sfm8568/qRb/yKkdCGECA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f400fc9b7648ce04653481c7506ad0c4c9c7a5f1d2be4210cfcf5dbc5cd9e5e9","last_reissued_at":"2026-07-05T00:47:20.014090Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:47:20.014090Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adv-BERT: BERT is not robust on misspellings! Generating nature adversarial samples on BERT","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Akari Asai, Caiming Xiong, Jia Li, Kazuma Hashimoto, Lichao Sun, Philip Yu, Wenpeng Yin","submitted_at":"2020-02-27T22:07:11Z","abstract_excerpt":"There is an increasing amount of literature that claims the brittleness of deep neural networks in dealing with adversarial examples that are created maliciously. It is unclear, however, how the models will perform in realistic scenarios where \\textit{natural rather than malicious} adversarial instances often exist. This work systematically explores the robustness of BERT, the state-of-the-art Transformer-style model in NLP, in dealing with noisy data, particularly mistakes in typing the keyboard, that occur inadvertently. Intensive experiments on sentiment analysis and question answering benc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2003.04985","kind":"arxiv","version":1},"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/2003.04985/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":"2003.04985","created_at":"2026-07-05T00:47:20.014150+00:00"},{"alias_kind":"arxiv_version","alias_value":"2003.04985v1","created_at":"2026-07-05T00:47:20.014150+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2003.04985","created_at":"2026-07-05T00:47:20.014150+00:00"},{"alias_kind":"pith_short_12","alias_value":"6QAPZG3WJDHA","created_at":"2026-07-05T00:47:20.014150+00:00"},{"alias_kind":"pith_short_16","alias_value":"6QAPZG3WJDHAIZJU","created_at":"2026-07-05T00:47:20.014150+00:00"},{"alias_kind":"pith_short_8","alias_value":"6QAPZG3W","created_at":"2026-07-05T00:47:20.014150+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.22202","citing_title":"Library Hallucinations in LLM-Generated Code: A Risk Analysis Grounded in Developer Queries","ref_index":60,"is_internal_anchor":false},{"citing_arxiv_id":"2412.05579","citing_title":"LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods","ref_index":218,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08044","citing_title":"Fast Byte Latent Transformer","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17105","citing_title":"How Tokenization Limits Phonological Knowledge Representation in Language Models and How to Improve Them","ref_index":45,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT","json":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT.json","graph_json":"https://pith.science/api/pith-number/6QAPZG3WJDHAIZJUQHDVA2WQYT/graph.json","events_json":"https://pith.science/api/pith-number/6QAPZG3WJDHAIZJUQHDVA2WQYT/events.json","paper":"https://pith.science/paper/6QAPZG3W"},"agent_actions":{"view_html":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT","download_json":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT.json","view_paper":"https://pith.science/paper/6QAPZG3W","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2003.04985&json=true","fetch_graph":"https://pith.science/api/pith-number/6QAPZG3WJDHAIZJUQHDVA2WQYT/graph.json","fetch_events":"https://pith.science/api/pith-number/6QAPZG3WJDHAIZJUQHDVA2WQYT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT/action/storage_attestation","attest_author":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT/action/author_attestation","sign_citation":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT/action/citation_signature","submit_replication":"https://pith.science/pith/6QAPZG3WJDHAIZJUQHDVA2WQYT/action/replication_record"}},"created_at":"2026-07-05T00:47:20.014150+00:00","updated_at":"2026-07-05T00:47:20.014150+00:00"}