{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:4UOLLLVX6HO4IS52ISA3MKTLZI","short_pith_number":"pith:4UOLLLVX","schema_version":"1.0","canonical_sha256":"e51cb5aeb7f1ddc44bba4481b62a6bca1ded80ef30fff449ad5c6cb1cdcb3e3a","source":{"kind":"arxiv","id":"1905.06922","version":1},"attestation_state":"computed","paper":{"title":"On Variational Bounds of Mutual Information","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaron van den Oord, Alexander A. Alemi, Ben Poole, George Tucker, Sherjil Ozair","submitted_at":"2019-05-16T17:31:53Z","abstract_excerpt":"Estimating and optimizing Mutual Information (MI) is core to many problems in machine learning; however, bounding MI in high dimensions is challenging. To establish tractable and scalable objectives, recent work has turned to variational bounds parameterized by neural networks, but the relationships and tradeoffs between these bounds remains unclear. In this work, we unify these recent developments in a single framework. We find that the existing variational lower bounds degrade when the MI is large, exhibiting either high bias or high variance. To address this problem, we introduce a continuu"},"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":"1905.06922","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-16T17:31:53Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"21937672b9b9b11de93dd98e3442c9862c00164e2d27939ec3992b84f4eb7182","abstract_canon_sha256":"4140f5432aa3bcaaace1197143ad281223e7d455a43200c94e1953db22ca9c9b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:45:58.525214Z","signature_b64":"xuvGP9a5fQgYT2IuM1y1jW24dDYwDClFqUU0z3tnpeHtLCCSqyMV1Riz8MmNwlNS6vZW+Zz2p8hTYZJGcUQjDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e51cb5aeb7f1ddc44bba4481b62a6bca1ded80ef30fff449ad5c6cb1cdcb3e3a","last_reissued_at":"2026-05-17T23:45:58.524566Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:45:58.524566Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On Variational Bounds of Mutual Information","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaron van den Oord, Alexander A. Alemi, Ben Poole, George Tucker, Sherjil Ozair","submitted_at":"2019-05-16T17:31:53Z","abstract_excerpt":"Estimating and optimizing Mutual Information (MI) is core to many problems in machine learning; however, bounding MI in high dimensions is challenging. To establish tractable and scalable objectives, recent work has turned to variational bounds parameterized by neural networks, but the relationships and tradeoffs between these bounds remains unclear. In this work, we unify these recent developments in a single framework. We find that the existing variational lower bounds degrade when the MI is large, exhibiting either high bias or high variance. To address this problem, we introduce a continuu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.06922","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":""},"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":"1905.06922","created_at":"2026-05-17T23:45:58.524663+00:00"},{"alias_kind":"arxiv_version","alias_value":"1905.06922v1","created_at":"2026-05-17T23:45:58.524663+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.06922","created_at":"2026-05-17T23:45:58.524663+00:00"},{"alias_kind":"pith_short_12","alias_value":"4UOLLLVX6HO4","created_at":"2026-05-18T12:33:10.108867+00:00"},{"alias_kind":"pith_short_16","alias_value":"4UOLLLVX6HO4IS52","created_at":"2026-05-18T12:33:10.108867+00:00"},{"alias_kind":"pith_short_8","alias_value":"4UOLLLVX","created_at":"2026-05-18T12:33:10.108867+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.11870","citing_title":"Information theoretic underpinning of self-supervised learning by clustering","ref_index":168,"is_internal_anchor":false},{"citing_arxiv_id":"1912.01603","citing_title":"Dream to Control: Learning Behaviors by Latent Imagination","ref_index":40,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI","json":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI.json","graph_json":"https://pith.science/api/pith-number/4UOLLLVX6HO4IS52ISA3MKTLZI/graph.json","events_json":"https://pith.science/api/pith-number/4UOLLLVX6HO4IS52ISA3MKTLZI/events.json","paper":"https://pith.science/paper/4UOLLLVX"},"agent_actions":{"view_html":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI","download_json":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI.json","view_paper":"https://pith.science/paper/4UOLLLVX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1905.06922&json=true","fetch_graph":"https://pith.science/api/pith-number/4UOLLLVX6HO4IS52ISA3MKTLZI/graph.json","fetch_events":"https://pith.science/api/pith-number/4UOLLLVX6HO4IS52ISA3MKTLZI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI/action/storage_attestation","attest_author":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI/action/author_attestation","sign_citation":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI/action/citation_signature","submit_replication":"https://pith.science/pith/4UOLLLVX6HO4IS52ISA3MKTLZI/action/replication_record"}},"created_at":"2026-05-17T23:45:58.524663+00:00","updated_at":"2026-05-17T23:45:58.524663+00:00"}