{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:OEDSR3KU7MJMLWD5B7UVRP6E4S","short_pith_number":"pith:OEDSR3KU","schema_version":"1.0","canonical_sha256":"710728ed54fb12c5d87d0fe958bfc4e4b275178848a4fff943d698d213479daa","source":{"kind":"arxiv","id":"2004.07683","version":2},"attestation_state":"computed","paper":{"title":"Do sequence-to-sequence VAEs learn global features of sentences?","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Pascal Vincent, Tom Bosc","submitted_at":"2020-04-16T14:43:27Z","abstract_excerpt":"Autoregressive language models are powerful and relatively easy to train. However, these models are usually trained without explicit conditioning labels and do not offer easy ways to control global aspects such as sentiment or topic during generation. Bowman & al. (2016) adapted the Variational Autoencoder (VAE) for natural language with the sequence-to-sequence architecture and claimed that the latent vector was able to capture such global features in an unsupervised manner. We question this claim. We measure which words benefit most from the latent information by decomposing the reconstructi"},"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":"2004.07683","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2020-04-16T14:43:27Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"9e5204a41582954dcec2334ab6125eec9f35e094427f42e9430ec213a76bae3e","abstract_canon_sha256":"e4e7476e981c7e27d3eb34a2d134b88a1f3ea2b0c1f53bcfd21063a12f68d192"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:26:47.752219Z","signature_b64":"YEkrBaFjwlAuva5sdxO/cSptH69yp7XJx6JAA3W/OKcJ2kBgDIpVkLoyyV7wsNn0wCVtx5XLUI2Jirxvh64QBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"710728ed54fb12c5d87d0fe958bfc4e4b275178848a4fff943d698d213479daa","last_reissued_at":"2026-07-05T02:26:47.751512Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:26:47.751512Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do sequence-to-sequence VAEs learn global features of sentences?","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Pascal Vincent, Tom Bosc","submitted_at":"2020-04-16T14:43:27Z","abstract_excerpt":"Autoregressive language models are powerful and relatively easy to train. However, these models are usually trained without explicit conditioning labels and do not offer easy ways to control global aspects such as sentiment or topic during generation. Bowman & al. (2016) adapted the Variational Autoencoder (VAE) for natural language with the sequence-to-sequence architecture and claimed that the latent vector was able to capture such global features in an unsupervised manner. We question this claim. We measure which words benefit most from the latent information by decomposing the reconstructi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2004.07683","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/2004.07683/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":"2004.07683","created_at":"2026-07-05T02:26:47.751601+00:00"},{"alias_kind":"arxiv_version","alias_value":"2004.07683v2","created_at":"2026-07-05T02:26:47.751601+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2004.07683","created_at":"2026-07-05T02:26:47.751601+00:00"},{"alias_kind":"pith_short_12","alias_value":"OEDSR3KU7MJM","created_at":"2026-07-05T02:26:47.751601+00:00"},{"alias_kind":"pith_short_16","alias_value":"OEDSR3KU7MJMLWD5","created_at":"2026-07-05T02:26:47.751601+00:00"},{"alias_kind":"pith_short_8","alias_value":"OEDSR3KU","created_at":"2026-07-05T02:26:47.751601+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/OEDSR3KU7MJMLWD5B7UVRP6E4S","json":"https://pith.science/pith/OEDSR3KU7MJMLWD5B7UVRP6E4S.json","graph_json":"https://pith.science/api/pith-number/OEDSR3KU7MJMLWD5B7UVRP6E4S/graph.json","events_json":"https://pith.science/api/pith-number/OEDSR3KU7MJMLWD5B7UVRP6E4S/events.json","paper":"https://pith.science/paper/OEDSR3KU"},"agent_actions":{"view_html":"https://pith.science/pith/OEDSR3KU7MJMLWD5B7UVRP6E4S","download_json":"https://pith.science/pith/OEDSR3KU7MJMLWD5B7UVRP6E4S.json","view_paper":"https://pith.science/paper/OEDSR3KU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2004.07683&json=true","fetch_graph":"https://pith.science/api/pith-number/OEDSR3KU7MJMLWD5B7UVRP6E4S/graph.json","fetch_events":"https://pith.science/api/pith-number/OEDSR3KU7MJMLWD5B7UVRP6E4S/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OEDSR3KU7MJMLWD5B7UVRP6E4S/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OEDSR3KU7MJMLWD5B7UVRP6E4S/action/storage_attestation","attest_author":"https://pith.science/pith/OEDSR3KU7MJMLWD5B7UVRP6E4S/action/author_attestation","sign_citation":"https://pith.science/pith/OEDSR3KU7MJMLWD5B7UVRP6E4S/action/citation_signature","submit_replication":"https://pith.science/pith/OEDSR3KU7MJMLWD5B7UVRP6E4S/action/replication_record"}},"created_at":"2026-07-05T02:26:47.751601+00:00","updated_at":"2026-07-05T02:26:47.751601+00:00"}