{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:KE6NQEW6ICJGUYIRXVA3VN646P","short_pith_number":"pith:KE6NQEW6","schema_version":"1.0","canonical_sha256":"513cd812de40926a6111bd41bab7dcf3c2e9b98f5c3e4189b72e99add166f824","source":{"kind":"arxiv","id":"2107.02331","version":1},"attestation_state":"computed","paper":{"title":"Mind Your Outliers! Investigating the Negative Impact of Outliers on Active Learning for Visual Question Answering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Christopher D. Manning, Li Fei-Fei, Ranjay Krishna, Siddharth Karamcheti","submitted_at":"2021-07-06T00:52:11Z","abstract_excerpt":"Active learning promises to alleviate the massive data needs of supervised machine learning: it has successfully improved sample efficiency by an order of magnitude on traditional tasks like topic classification and object recognition. However, we uncover a striking contrast to this promise: across 5 models and 4 datasets on the task of visual question answering, a wide variety of active learning approaches fail to outperform random selection. To understand this discrepancy, we profile 8 active learning methods on a per-example basis, and identify the problem as collective outliers -- groups o"},"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":"2107.02331","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-07-06T00:52:11Z","cross_cats_sorted":["cs.AI","cs.CV","cs.LG"],"title_canon_sha256":"dff51ab47a89d572bb6adb12a237dcebb9cbe66f61c8288e87cb7f73793f481a","abstract_canon_sha256":"2d58cd890b9cf5deb7d38fced5035deb910d5787eb9fc0a74ebb0d9d7b5772d4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:55:19.372902Z","signature_b64":"JuRQ2/Ij8GjEjZhWCSQRAIP9Qv2aDjcfwQ2GhC5/mkKP3klkFoQOtohClojDGNNt0GwONE2Evfh2OmuIfH9xAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"513cd812de40926a6111bd41bab7dcf3c2e9b98f5c3e4189b72e99add166f824","last_reissued_at":"2026-07-05T02:55:19.372445Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:55:19.372445Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mind Your Outliers! Investigating the Negative Impact of Outliers on Active Learning for Visual Question Answering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Christopher D. Manning, Li Fei-Fei, Ranjay Krishna, Siddharth Karamcheti","submitted_at":"2021-07-06T00:52:11Z","abstract_excerpt":"Active learning promises to alleviate the massive data needs of supervised machine learning: it has successfully improved sample efficiency by an order of magnitude on traditional tasks like topic classification and object recognition. However, we uncover a striking contrast to this promise: across 5 models and 4 datasets on the task of visual question answering, a wide variety of active learning approaches fail to outperform random selection. To understand this discrepancy, we profile 8 active learning methods on a per-example basis, and identify the problem as collective outliers -- groups o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2107.02331","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/2107.02331/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":"2107.02331","created_at":"2026-07-05T02:55:19.372509+00:00"},{"alias_kind":"arxiv_version","alias_value":"2107.02331v1","created_at":"2026-07-05T02:55:19.372509+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2107.02331","created_at":"2026-07-05T02:55:19.372509+00:00"},{"alias_kind":"pith_short_12","alias_value":"KE6NQEW6ICJG","created_at":"2026-07-05T02:55:19.372509+00:00"},{"alias_kind":"pith_short_16","alias_value":"KE6NQEW6ICJGUYIR","created_at":"2026-07-05T02:55:19.372509+00:00"},{"alias_kind":"pith_short_8","alias_value":"KE6NQEW6","created_at":"2026-07-05T02:55:19.372509+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.00474","citing_title":"ADAptation: Reconstruction-based Unsupervised Active Learning for Breast Ultrasound Diagnosis","ref_index":14,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P","json":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P.json","graph_json":"https://pith.science/api/pith-number/KE6NQEW6ICJGUYIRXVA3VN646P/graph.json","events_json":"https://pith.science/api/pith-number/KE6NQEW6ICJGUYIRXVA3VN646P/events.json","paper":"https://pith.science/paper/KE6NQEW6"},"agent_actions":{"view_html":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P","download_json":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P.json","view_paper":"https://pith.science/paper/KE6NQEW6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2107.02331&json=true","fetch_graph":"https://pith.science/api/pith-number/KE6NQEW6ICJGUYIRXVA3VN646P/graph.json","fetch_events":"https://pith.science/api/pith-number/KE6NQEW6ICJGUYIRXVA3VN646P/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P/action/storage_attestation","attest_author":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P/action/author_attestation","sign_citation":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P/action/citation_signature","submit_replication":"https://pith.science/pith/KE6NQEW6ICJGUYIRXVA3VN646P/action/replication_record"}},"created_at":"2026-07-05T02:55:19.372509+00:00","updated_at":"2026-07-05T02:55:19.372509+00:00"}