{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:JQUUQ5HS75TPEC3HF27ILT35BA","short_pith_number":"pith:JQUUQ5HS","schema_version":"1.0","canonical_sha256":"4c294874f2ff66f20b672ebe85cf7d080bd9a39a7566995edb9fadbf1726e13f","source":{"kind":"arxiv","id":"2210.10318","version":1},"attestation_state":"computed","paper":{"title":"Gaussian-Bernoulli RBMs Without Tears","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"David J. Fleet, Geoffrey Hinton, Mengye Ren, Renjie Liao, Simon Kornblith","submitted_at":"2022-10-19T06:22:55Z","abstract_excerpt":"We revisit the challenging problem of training Gaussian-Bernoulli restricted Boltzmann machines (GRBMs), introducing two innovations. We propose a novel Gibbs-Langevin sampling algorithm that outperforms existing methods like Gibbs sampling. We propose a modified contrastive divergence (CD) algorithm so that one can generate images with GRBMs starting from noise. This enables direct comparison of GRBMs with deep generative models, improving evaluation protocols in the RBM literature. Moreover, we show that modified CD and gradient clipping are enough to robustly train GRBMs with large learning"},"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":"2210.10318","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-10-19T06:22:55Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"9a2674f5846f81fc58af6bc85a5e257faee88c630604eee6c1cfaad0993dce4c","abstract_canon_sha256":"8fd4c40b742da7d73cb1f5f05ab3ac37aa08039e60819addd2f771eea7b876d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:08:08.738762Z","signature_b64":"IP6FFueYYKpfJI2hJ19PnDITtjBmNLC/fmR9xYTvGEppI82bfts25/WqjSq6jEQmgTbtfdqquuh7Z7lf5AaUBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4c294874f2ff66f20b672ebe85cf7d080bd9a39a7566995edb9fadbf1726e13f","last_reissued_at":"2026-07-05T05:08:08.738409Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:08:08.738409Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Gaussian-Bernoulli RBMs Without Tears","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"David J. Fleet, Geoffrey Hinton, Mengye Ren, Renjie Liao, Simon Kornblith","submitted_at":"2022-10-19T06:22:55Z","abstract_excerpt":"We revisit the challenging problem of training Gaussian-Bernoulli restricted Boltzmann machines (GRBMs), introducing two innovations. We propose a novel Gibbs-Langevin sampling algorithm that outperforms existing methods like Gibbs sampling. We propose a modified contrastive divergence (CD) algorithm so that one can generate images with GRBMs starting from noise. This enables direct comparison of GRBMs with deep generative models, improving evaluation protocols in the RBM literature. Moreover, we show that modified CD and gradient clipping are enough to robustly train GRBMs with large learning"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.10318","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/2210.10318/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":"2210.10318","created_at":"2026-07-05T05:08:08.738468+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.10318v1","created_at":"2026-07-05T05:08:08.738468+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.10318","created_at":"2026-07-05T05:08:08.738468+00:00"},{"alias_kind":"pith_short_12","alias_value":"JQUUQ5HS75TP","created_at":"2026-07-05T05:08:08.738468+00:00"},{"alias_kind":"pith_short_16","alias_value":"JQUUQ5HS75TPEC3H","created_at":"2026-07-05T05:08:08.738468+00:00"},{"alias_kind":"pith_short_8","alias_value":"JQUUQ5HS","created_at":"2026-07-05T05:08:08.738468+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2301.09428","citing_title":"Explaining the effects of non-convergent sampling in the training of Energy-Based Models","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2411.10406","citing_title":"How to Build a Quantum Supercomputer: Scaling from Hundreds to Millions of Qubits","ref_index":162,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA","json":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA.json","graph_json":"https://pith.science/api/pith-number/JQUUQ5HS75TPEC3HF27ILT35BA/graph.json","events_json":"https://pith.science/api/pith-number/JQUUQ5HS75TPEC3HF27ILT35BA/events.json","paper":"https://pith.science/paper/JQUUQ5HS"},"agent_actions":{"view_html":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA","download_json":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA.json","view_paper":"https://pith.science/paper/JQUUQ5HS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.10318&json=true","fetch_graph":"https://pith.science/api/pith-number/JQUUQ5HS75TPEC3HF27ILT35BA/graph.json","fetch_events":"https://pith.science/api/pith-number/JQUUQ5HS75TPEC3HF27ILT35BA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA/action/storage_attestation","attest_author":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA/action/author_attestation","sign_citation":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA/action/citation_signature","submit_replication":"https://pith.science/pith/JQUUQ5HS75TPEC3HF27ILT35BA/action/replication_record"}},"created_at":"2026-07-05T05:08:08.738468+00:00","updated_at":"2026-07-05T05:08:08.738468+00:00"}