{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:N6J2QMOJJZPVT6BFBVRVJR23TN","short_pith_number":"pith:N6J2QMOJ","schema_version":"1.0","canonical_sha256":"6f93a831c94e5f59f8250d6354c75b9b50a36a957b0e27c1887bfb73f71f6083","source":{"kind":"arxiv","id":"1912.02757","version":2},"attestation_state":"computed","paper":{"title":"Deep Ensembles: A Loss Landscape Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Balaji Lakshminarayanan, Huiyi Hu, Stanislav Fort","submitted_at":"2019-12-05T17:48:18Z","abstract_excerpt":"Deep ensembles have been empirically shown to be a promising approach for improving accuracy, uncertainty and out-of-distribution robustness of deep learning models. While deep ensembles were theoretically motivated by the bootstrap, non-bootstrap ensembles trained with just random initialization also perform well in practice, which suggests that there could be other explanations for why deep ensembles work well. Bayesian neural networks, which learn distributions over the parameters of the network, are theoretically well-motivated by Bayesian principles, but do not perform as well as deep ens"},"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":"1912.02757","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2019-12-05T17:48:18Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f6d0646135a13d9c1469fb94da047c9666b204b37d9db9c6256c4a3df5cb63c5","abstract_canon_sha256":"bc362d20d2cf1eced7dcd15d489c8c430674700ddaad7ba519626890751d1732"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:13:24.164288Z","signature_b64":"BNSDUB3IlRzeD6uPzgThx5UUL/dYjBoy2pc3jrnpwjwawGKfMQkc5cdyEpkkMJvE7P5zoBLu2ve+fAv2jTxjDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6f93a831c94e5f59f8250d6354c75b9b50a36a957b0e27c1887bfb73f71f6083","last_reissued_at":"2026-07-05T01:13:24.163743Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:13:24.163743Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Deep Ensembles: A Loss Landscape Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Balaji Lakshminarayanan, Huiyi Hu, Stanislav Fort","submitted_at":"2019-12-05T17:48:18Z","abstract_excerpt":"Deep ensembles have been empirically shown to be a promising approach for improving accuracy, uncertainty and out-of-distribution robustness of deep learning models. While deep ensembles were theoretically motivated by the bootstrap, non-bootstrap ensembles trained with just random initialization also perform well in practice, which suggests that there could be other explanations for why deep ensembles work well. Bayesian neural networks, which learn distributions over the parameters of the network, are theoretically well-motivated by Bayesian principles, but do not perform as well as deep ens"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1912.02757","kind":"arxiv","version":2},"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/1912.02757/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":"1912.02757","created_at":"2026-07-05T01:13:24.163802+00:00"},{"alias_kind":"arxiv_version","alias_value":"1912.02757v2","created_at":"2026-07-05T01:13:24.163802+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1912.02757","created_at":"2026-07-05T01:13:24.163802+00:00"},{"alias_kind":"pith_short_12","alias_value":"N6J2QMOJJZPV","created_at":"2026-07-05T01:13:24.163802+00:00"},{"alias_kind":"pith_short_16","alias_value":"N6J2QMOJJZPVT6BF","created_at":"2026-07-05T01:13:24.163802+00:00"},{"alias_kind":"pith_short_8","alias_value":"N6J2QMOJ","created_at":"2026-07-05T01:13:24.163802+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":23,"internal_anchor_count":4,"sample":[{"citing_arxiv_id":"2607.06628","citing_title":"Cross-Trajectory Chimera Interventions Reveal Dissociable Roles of Weight Magnitude and Direction in Grokking","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2607.06776","citing_title":"Efficient Bayesian Deep Ensembles via Analytic Predictive Inference","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2607.08493","citing_title":"Ensemble Diversity Optimization for Subjective Supervision","ref_index":74,"is_internal_anchor":true},{"citing_arxiv_id":"2607.05380","citing_title":"TabPack: Efficient Hyperparameter Ensembles for Tabular Deep Learning","ref_index":136,"is_internal_anchor":true},{"citing_arxiv_id":"2606.20536","citing_title":"The FID Lottery: Quantifying Hidden Randomness in Generative-Model Evaluation","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10929","citing_title":"Recoverable but Not Stationary:Local Linear Structures in Weights and Activations","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03938","citing_title":"q0: Primitives for Hyper-Epoch Pretraining","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24340","citing_title":"ChainzRule: Sample-Efficient, Robust Deep Learning Across Tabular, NLP, and Vision Tasks","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27747","citing_title":"Soft Specialists: $\\alpha$-R\\'enyi Ensembles for Uncertainty-Aware LLM Post-Training","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30135","citing_title":"DAMEL: Dual-Axis Multi-Expert Learning for Class-Imbalanced Learning","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23315","citing_title":"Convergence Without Understanding: When Language Models Agree on Representations but Disagree on Reasoning","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2504.19239","citing_title":"The effect of the number of parameters and the number of local feature patches on loss landscapes in distributed quantum neural networks","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22390","citing_title":"A Posterior-Predictive Variance Decomposition for Epistemic and Aleatoric Uncertainty in Wind Power Forecasting","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22593","citing_title":"Do Deep Ensembles Actually Capture Uncertainty in Graph Neural Networks?","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2603.03190","citing_title":"Expectation and Acoustic Neural Network Representations Enhance Music Identification from Brain Activity","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20341","citing_title":"Causal Unlearning in Collaborative Optimization: Exact and Approximate Influence Reversal under Adversarial Contributions","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18472","citing_title":"Flowing with Confidence","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04038","citing_title":"FLAME: Condensing Ensemble Diversity into a Single Network for Efficient Sequential Recommendation","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2212.04089","citing_title":"Editing Models with Task Arithmetic","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24708","citing_title":"Scalable Hyperparameter-Divergent Ensemble Training with Automatic Learning Rate Exploration for Large Models","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07244","citing_title":"Experience Sharing in Mutual Reinforcement Learning for Heterogeneous Language Models","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07963","citing_title":"Rethinking Data Mixing from the Perspective of Large Language Models","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13992","citing_title":"Physics-Informed Neural Networks for Methane Sorption: Cross-Gas Transfer Learning, Ensemble Collapse Under Physics Constraints, and Monte Carlo Dropout Uncertainty Quantification","ref_index":78,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN","json":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN.json","graph_json":"https://pith.science/api/pith-number/N6J2QMOJJZPVT6BFBVRVJR23TN/graph.json","events_json":"https://pith.science/api/pith-number/N6J2QMOJJZPVT6BFBVRVJR23TN/events.json","paper":"https://pith.science/paper/N6J2QMOJ"},"agent_actions":{"view_html":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN","download_json":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN.json","view_paper":"https://pith.science/paper/N6J2QMOJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1912.02757&json=true","fetch_graph":"https://pith.science/api/pith-number/N6J2QMOJJZPVT6BFBVRVJR23TN/graph.json","fetch_events":"https://pith.science/api/pith-number/N6J2QMOJJZPVT6BFBVRVJR23TN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN/action/storage_attestation","attest_author":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN/action/author_attestation","sign_citation":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN/action/citation_signature","submit_replication":"https://pith.science/pith/N6J2QMOJJZPVT6BFBVRVJR23TN/action/replication_record"}},"created_at":"2026-07-05T01:13:24.163802+00:00","updated_at":"2026-07-05T01:13:24.163802+00:00"}