{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2018:D5G3NZC5XQ34TCFLIDJI2F3GKH","short_pith_number":"pith:D5G3NZC5","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"},"canonical_sha256":"1f4db6e45dbc37c988ab40d28d176651da68f69ad257a0e2e7e1004447118498","source":{"kind":"arxiv","id":"1812.04616","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1812.04616","created_at":"2026-05-17T23:50:40Z"},{"alias_kind":"arxiv_version","alias_value":"1812.04616v3","created_at":"2026-05-17T23:50:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1812.04616","created_at":"2026-05-17T23:50:40Z"},{"alias_kind":"pith_short_12","alias_value":"D5G3NZC5XQ34","created_at":"2026-05-18T12:32:19Z"},{"alias_kind":"pith_short_16","alias_value":"D5G3NZC5XQ34TCFL","created_at":"2026-05-18T12:32:19Z"},{"alias_kind":"pith_short_8","alias_value":"D5G3NZC5","created_at":"2026-05-18T12:32:19Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2018:D5G3NZC5XQ34TCFLIDJI2F3GKH","target":"record","payload":{"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"},"canonical_sha256":"1f4db6e45dbc37c988ab40d28d176651da68f69ad257a0e2e7e1004447118498","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"},"source_kind":"arxiv","source_id":"1812.04616","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:50:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"zJTd6pJMlG3cVH5hl9MbVhL9mHx1kaG3h+PA7faSyS0jmXLVZK+RLlnehsatwd78n7EVmh+32G72snvxKu1XDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T14:53:01.767689Z"},"content_sha256":"881fbf903b2305f034f7663bb9d12e6592400b005128e3a4e370baee57fb3b1c","schema_version":"1.0","event_id":"sha256:881fbf903b2305f034f7663bb9d12e6592400b005128e3a4e370baee57fb3b1c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2018:D5G3NZC5XQ34TCFLIDJI2F3GKH","target":"graph","payload":{"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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:50:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Tw5H2BtBY3BPjy01KiDV/n+1ABsW1cSFzgpZlG/x4o5PCkk9bMYmaePb9qbo9NeVC8hKPNYiTL/cOP9EI8YwDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T14:53:01.768040Z"},"content_sha256":"f3dbba38cfb3c0c260c2f5f2542e2bbdfdd4fc70acc46beb218f155075e5f132","schema_version":"1.0","event_id":"sha256:f3dbba38cfb3c0c260c2f5f2542e2bbdfdd4fc70acc46beb218f155075e5f132"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/bundle.json","state_url":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-03T14:53:01Z","links":{"resolver":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH","bundle":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/bundle.json","state":"https://pith.science/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/state.json","well_known_bundle":"https://pith.science/.well-known/pith/D5G3NZC5XQ34TCFLIDJI2F3GKH/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2018:D5G3NZC5XQ34TCFLIDJI2F3GKH","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"9b093e0fbf53999b9599238314829cf6ec6399486b18c13803b1a7c4173b1c3b","cross_cats_sorted":["cs.LG","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2018-12-10T20:00:36Z","title_canon_sha256":"cdd6e50909e14246843450d58e41a4ebce35c8fe4da2af0e099601ca4de75aa6"},"schema_version":"1.0","source":{"id":"1812.04616","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1812.04616","created_at":"2026-05-17T23:50:40Z"},{"alias_kind":"arxiv_version","alias_value":"1812.04616v3","created_at":"2026-05-17T23:50:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1812.04616","created_at":"2026-05-17T23:50:40Z"},{"alias_kind":"pith_short_12","alias_value":"D5G3NZC5XQ34","created_at":"2026-05-18T12:32:19Z"},{"alias_kind":"pith_short_16","alias_value":"D5G3NZC5XQ34TCFL","created_at":"2026-05-18T12:32:19Z"},{"alias_kind":"pith_short_8","alias_value":"D5G3NZC5","created_at":"2026-05-18T12:32:19Z"}],"graph_snapshots":[{"event_id":"sha256:f3dbba38cfb3c0c260c2f5f2542e2bbdfdd4fc70acc46beb218f155075e5f132","target":"graph","created_at":"2026-05-17T23:50:40Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"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 ","authors_text":"Sachin Kumar, Yulia Tsvetkov","cross_cats":["cs.LG","stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2018-12-10T20:00:36Z","title":"Von Mises-Fisher Loss for Training Sequence to Sequence Models with Continuous Outputs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1812.04616","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:881fbf903b2305f034f7663bb9d12e6592400b005128e3a4e370baee57fb3b1c","target":"record","created_at":"2026-05-17T23:50:40Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"9b093e0fbf53999b9599238314829cf6ec6399486b18c13803b1a7c4173b1c3b","cross_cats_sorted":["cs.LG","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2018-12-10T20:00:36Z","title_canon_sha256":"cdd6e50909e14246843450d58e41a4ebce35c8fe4da2af0e099601ca4de75aa6"},"schema_version":"1.0","source":{"id":"1812.04616","kind":"arxiv","version":3}},"canonical_sha256":"1f4db6e45dbc37c988ab40d28d176651da68f69ad257a0e2e7e1004447118498","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1f4db6e45dbc37c988ab40d28d176651da68f69ad257a0e2e7e1004447118498","first_computed_at":"2026-05-17T23:50:40.825542Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:50:40.825542Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Ptvr4HMTP1FG4gEJJKV0Mnl+IlTa9dP4QbRJKlBT2OZvAmCDBDE4TvWvsMosS2emVf5dEoP6UPgdQcrSyL72Cg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:50:40.825949Z","signed_message":"canonical_sha256_bytes"},"source_id":"1812.04616","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:881fbf903b2305f034f7663bb9d12e6592400b005128e3a4e370baee57fb3b1c","sha256:f3dbba38cfb3c0c260c2f5f2542e2bbdfdd4fc70acc46beb218f155075e5f132"],"state_sha256":"da900bdfd79f6e9fce66620cbda9a00c3150832a00ee9ab2e8f5ce823c06857b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KZ/VhqkAOkDyCFu4AEBG4yhVXlmcb+pyf+WbWnqkQJPcbHfI/EKOsXuNIGiRFrgGu0aWI6EPPgyf4DlcCpRCDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-03T14:53:01.770065Z","bundle_sha256":"e656b30ecc15e13868ad6f59f6cd35d54b3dc33237c0066b32a2e0e424e431d2"}}