{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:D5G3NZC5XQ34TCFLIDJI2F3GKH","short_pith_number":"pith:D5G3NZC5","schema_version":"1.0","canonical_sha256":"1f4db6e45dbc37c988ab40d28d176651da68f69ad257a0e2e7e1004447118498","source":{"kind":"arxiv","id":"1812.04616","version":3},"attestation_state":"computed","paper":{"title":"Von Mises-Fisher Loss for Training Sequence to Sequence Models with Continuous Outputs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Sachin Kumar, Yulia Tsvetkov","submitted_at":"2018-12-10T20:00:36Z","abstract_excerpt":"The Softmax function is used in the final layer of nearly all existing sequence-to-sequence models for language generation. However, it is usually the slowest layer to compute which limits the vocabulary size to a subset of most frequent types; and it has a large memory footprint. We propose a general technique for replacing the softmax layer with a continuous embedding layer. Our primary innovations are a novel probabilistic loss, and a training and inference procedure in which we generate a probability distribution over pre-trained word embeddings, instead of a multinomial distribution over "},"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":"1812.04616","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2018-12-10T20:00:36Z","cross_cats_sorted":["cs.LG","stat.ML"],"title_canon_sha256":"cdd6e50909e14246843450d58e41a4ebce35c8fe4da2af0e099601ca4de75aa6","abstract_canon_sha256":"9b093e0fbf53999b9599238314829cf6ec6399486b18c13803b1a7c4173b1c3b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:50:40.825949Z","signature_b64":"Ptvr4HMTP1FG4gEJJKV0Mnl+IlTa9dP4QbRJKlBT2OZvAmCDBDE4TvWvsMosS2emVf5dEoP6UPgdQcrSyL72Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1f4db6e45dbc37c988ab40d28d176651da68f69ad257a0e2e7e1004447118498","last_reissued_at":"2026-05-17T23:50:40.825542Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:50:40.825542Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Von Mises-Fisher Loss for Training Sequence to Sequence Models with Continuous Outputs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Sachin Kumar, Yulia Tsvetkov","submitted_at":"2018-12-10T20:00:36Z","abstract_excerpt":"The Softmax function is used in the final layer of nearly all existing sequence-to-sequence models for language generation. However, it is usually the slowest layer to compute which limits the vocabulary size to a subset of most frequent types; and it has a large memory footprint. We propose a general technique for replacing the softmax layer with a continuous embedding layer. Our primary innovations are a novel probabilistic loss, and a training and inference procedure in which we generate a probability distribution over pre-trained word embeddings, instead of a multinomial distribution over "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1812.04616","kind":"arxiv","version":3},"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":"1812.04616","created_at":"2026-05-17T23:50:40.825602+00:00"},{"alias_kind":"arxiv_version","alias_value":"1812.04616v3","created_at":"2026-05-17T23:50:40.825602+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1812.04616","created_at":"2026-05-17T23:50:40.825602+00:00"},{"alias_kind":"pith_short_12","alias_value":"D5G3NZC5XQ34","created_at":"2026-05-18T12:32:19.392346+00:00"},{"alias_kind":"pith_short_16","alias_value":"D5G3NZC5XQ34TCFL","created_at":"2026-05-18T12:32:19.392346+00:00"},{"alias_kind":"pith_short_8","alias_value":"D5G3NZC5","created_at":"2026-05-18T12:32:19.392346+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/D5G3NZC5XQ34TCFLIDJI2F3GKH","json":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH.json","graph_json":"https://pith.science/api/pith-number/D5G3NZC5XQ34TCFLIDJI2F3GKH/graph.json","events_json":"https://pith.science/api/pith-number/D5G3NZC5XQ34TCFLIDJI2F3GKH/events.json","paper":"https://pith.science/paper/D5G3NZC5"},"agent_actions":{"view_html":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH","download_json":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH.json","view_paper":"https://pith.science/paper/D5G3NZC5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1812.04616&json=true","fetch_graph":"https://pith.science/api/pith-number/D5G3NZC5XQ34TCFLIDJI2F3GKH/graph.json","fetch_events":"https://pith.science/api/pith-number/D5G3NZC5XQ34TCFLIDJI2F3GKH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/action/storage_attestation","attest_author":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/action/author_attestation","sign_citation":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/action/citation_signature","submit_replication":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/action/replication_record"}},"created_at":"2026-05-17T23:50:40.825602+00:00","updated_at":"2026-05-17T23:50:40.825602+00:00"}