{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:5NCB22SZHPSM2F262RSALLY6GI","short_pith_number":"pith:5NCB22SZ","schema_version":"1.0","canonical_sha256":"eb441d6a593be4cd175ed46405af1e323d0940509b9758cdaa18bda0e326078c","source":{"kind":"arxiv","id":"2207.00137","version":1},"attestation_state":"computed","paper":{"title":"Robustness of Epinets against Distributional Shifts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Benjamin Van Roy, Ian Osband, Seyed Mohammad Asghari, Sven Gowal, Vikranth Dwaracherla, Xiuyuan Lu, Zheng Wen","submitted_at":"2022-07-01T01:07:11Z","abstract_excerpt":"Recent work introduced the epinet as a new approach to uncertainty modeling in deep learning. An epinet is a small neural network added to traditional neural networks, which, together, can produce predictive distributions. In particular, using an epinet can greatly improve the quality of joint predictions across multiple inputs, a measure of how well a neural network knows what it does not know. In this paper, we examine whether epinets can offer similar advantages under distributional shifts. We find that, across ImageNet-A/O/C, epinets generally improve robustness metrics. Moreover, these im"},"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":"2207.00137","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-07-01T01:07:11Z","cross_cats_sorted":[],"title_canon_sha256":"34fa312c2c486baec790d12c2396ecc0c7b55416e055e3ab7793ba7a3a969e1e","abstract_canon_sha256":"7f11622089345781bfcb38073180b802f53f29d1c7e9cbe506f81b729fc5b300"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:36:39.581996Z","signature_b64":"KaSmMMydZnSlPsnAur530GkP704FxZWU6ZOycFQCjWRY4k6C4VXy+G+sja2pYmX4d2CgadOdOKOyOEzqAblRDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eb441d6a593be4cd175ed46405af1e323d0940509b9758cdaa18bda0e326078c","last_reissued_at":"2026-07-05T04:36:39.581579Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:36:39.581579Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Robustness of Epinets against Distributional Shifts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Benjamin Van Roy, Ian Osband, Seyed Mohammad Asghari, Sven Gowal, Vikranth Dwaracherla, Xiuyuan Lu, Zheng Wen","submitted_at":"2022-07-01T01:07:11Z","abstract_excerpt":"Recent work introduced the epinet as a new approach to uncertainty modeling in deep learning. An epinet is a small neural network added to traditional neural networks, which, together, can produce predictive distributions. In particular, using an epinet can greatly improve the quality of joint predictions across multiple inputs, a measure of how well a neural network knows what it does not know. In this paper, we examine whether epinets can offer similar advantages under distributional shifts. We find that, across ImageNet-A/O/C, epinets generally improve robustness metrics. Moreover, these im"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2207.00137","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/2207.00137/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":"2207.00137","created_at":"2026-07-05T04:36:39.581639+00:00"},{"alias_kind":"arxiv_version","alias_value":"2207.00137v1","created_at":"2026-07-05T04:36:39.581639+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2207.00137","created_at":"2026-07-05T04:36:39.581639+00:00"},{"alias_kind":"pith_short_12","alias_value":"5NCB22SZHPSM","created_at":"2026-07-05T04:36:39.581639+00:00"},{"alias_kind":"pith_short_16","alias_value":"5NCB22SZHPSM2F26","created_at":"2026-07-05T04:36:39.581639+00:00"},{"alias_kind":"pith_short_8","alias_value":"5NCB22SZ","created_at":"2026-07-05T04:36:39.581639+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/5NCB22SZHPSM2F262RSALLY6GI","json":"https://pith.science/pith/5NCB22SZHPSM2F262RSALLY6GI.json","graph_json":"https://pith.science/api/pith-number/5NCB22SZHPSM2F262RSALLY6GI/graph.json","events_json":"https://pith.science/api/pith-number/5NCB22SZHPSM2F262RSALLY6GI/events.json","paper":"https://pith.science/paper/5NCB22SZ"},"agent_actions":{"view_html":"https://pith.science/pith/5NCB22SZHPSM2F262RSALLY6GI","download_json":"https://pith.science/pith/5NCB22SZHPSM2F262RSALLY6GI.json","view_paper":"https://pith.science/paper/5NCB22SZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2207.00137&json=true","fetch_graph":"https://pith.science/api/pith-number/5NCB22SZHPSM2F262RSALLY6GI/graph.json","fetch_events":"https://pith.science/api/pith-number/5NCB22SZHPSM2F262RSALLY6GI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5NCB22SZHPSM2F262RSALLY6GI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5NCB22SZHPSM2F262RSALLY6GI/action/storage_attestation","attest_author":"https://pith.science/pith/5NCB22SZHPSM2F262RSALLY6GI/action/author_attestation","sign_citation":"https://pith.science/pith/5NCB22SZHPSM2F262RSALLY6GI/action/citation_signature","submit_replication":"https://pith.science/pith/5NCB22SZHPSM2F262RSALLY6GI/action/replication_record"}},"created_at":"2026-07-05T04:36:39.581639+00:00","updated_at":"2026-07-05T04:36:39.581639+00:00"}