{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2017:MEFYDUGR2L6ZEHUVJCZ6O63CHE","short_pith_number":"pith:MEFYDUGR","canonical_record":{"source":{"id":"1709.09346","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-09-27T06:04:03Z","cross_cats_sorted":[],"title_canon_sha256":"68a7063ae097841732924d19cc12f11bd7d80b1650ad192ce9dd72a37b6f070a","abstract_canon_sha256":"64f16bd801fa75da84b2a8469cf020371ca21730d6a2da4aca40ebb826c31f0a"},"schema_version":"1.0"},"canonical_sha256":"610b81d0d1d2fd921e9548b3e77b62390718091946bca5cfec4a1237943ec293","source":{"kind":"arxiv","id":"1709.09346","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1709.09346","created_at":"2026-05-18T00:32:54Z"},{"alias_kind":"arxiv_version","alias_value":"1709.09346v2","created_at":"2026-05-18T00:32:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1709.09346","created_at":"2026-05-18T00:32:54Z"},{"alias_kind":"pith_short_12","alias_value":"MEFYDUGR2L6Z","created_at":"2026-05-18T12:31:31Z"},{"alias_kind":"pith_short_16","alias_value":"MEFYDUGR2L6ZEHUV","created_at":"2026-05-18T12:31:31Z"},{"alias_kind":"pith_short_8","alias_value":"MEFYDUGR","created_at":"2026-05-18T12:31:31Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2017:MEFYDUGR2L6ZEHUVJCZ6O63CHE","target":"record","payload":{"canonical_record":{"source":{"id":"1709.09346","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-09-27T06:04:03Z","cross_cats_sorted":[],"title_canon_sha256":"68a7063ae097841732924d19cc12f11bd7d80b1650ad192ce9dd72a37b6f070a","abstract_canon_sha256":"64f16bd801fa75da84b2a8469cf020371ca21730d6a2da4aca40ebb826c31f0a"},"schema_version":"1.0"},"canonical_sha256":"610b81d0d1d2fd921e9548b3e77b62390718091946bca5cfec4a1237943ec293","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:32:54.760653Z","signature_b64":"esce+84sdx2OMBspHO5GXR9ST+NwVggDR9TVEMlZqxUTX8EabIo65eft3oXnenCL6Y0a+gQdauOvJirjiPZbBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"610b81d0d1d2fd921e9548b3e77b62390718091946bca5cfec4a1237943ec293","last_reissued_at":"2026-05-18T00:32:54.759928Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:32:54.759928Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1709.09346","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:32:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ibOtqiUXlvIsYfag2WfxWIABwIHPG3g8+9HoGZtUGNIBIo+jrIRBeNGuhqP09i6I/KmmzxzEp5TOZ0ElcllhCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T12:21:21.959181Z"},"content_sha256":"e171c41d11d64b6fe392b488d5604fc9c5423ad66548a6251d20d1c5a27eca51","schema_version":"1.0","event_id":"sha256:e171c41d11d64b6fe392b488d5604fc9c5423ad66548a6251d20d1c5a27eca51"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2017:MEFYDUGR2L6ZEHUVJCZ6O63CHE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Cold-Start Reinforcement Learning with Softmax Policy Gradient","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Nan Ding, Radu Soricut","submitted_at":"2017-09-27T06:04:03Z","abstract_excerpt":"Policy-gradient approaches to reinforcement learning have two common and undesirable overhead procedures, namely warm-start training and sample variance reduction. In this paper, we describe a reinforcement learning method based on a softmax value function that requires neither of these procedures. Our method combines the advantages of policy-gradient methods with the efficiency and simplicity of maximum-likelihood approaches. We apply this new cold-start reinforcement learning method in training sequence generation models for structured output prediction problems. Empirical evidence validates"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1709.09346","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":""},"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":"2026-05-18T00:32:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ex/bNZLMV98AnFKEtNKqv/AF5jFecgEYJXJDnS8wGRGTWo3dTkxb6U6xoZCDElqbDuSZYYWlzVvlb1YSYFgWDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T12:21:21.959876Z"},"content_sha256":"94d76b7796d50cc3e8a2dfa9d33d634243500d513e35417d104709161195a643","schema_version":"1.0","event_id":"sha256:94d76b7796d50cc3e8a2dfa9d33d634243500d513e35417d104709161195a643"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/MEFYDUGR2L6ZEHUVJCZ6O63CHE/bundle.json","state_url":"https://pith.science/pith/MEFYDUGR2L6ZEHUVJCZ6O63CHE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/MEFYDUGR2L6ZEHUVJCZ6O63CHE/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":"2026-06-03T12:21:21Z","links":{"resolver":"https://pith.science/pith/MEFYDUGR2L6ZEHUVJCZ6O63CHE","bundle":"https://pith.science/pith/MEFYDUGR2L6ZEHUVJCZ6O63CHE/bundle.json","state":"https://pith.science/pith/MEFYDUGR2L6ZEHUVJCZ6O63CHE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/MEFYDUGR2L6ZEHUVJCZ6O63CHE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2017:MEFYDUGR2L6ZEHUVJCZ6O63CHE","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":"64f16bd801fa75da84b2a8469cf020371ca21730d6a2da4aca40ebb826c31f0a","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-09-27T06:04:03Z","title_canon_sha256":"68a7063ae097841732924d19cc12f11bd7d80b1650ad192ce9dd72a37b6f070a"},"schema_version":"1.0","source":{"id":"1709.09346","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1709.09346","created_at":"2026-05-18T00:32:54Z"},{"alias_kind":"arxiv_version","alias_value":"1709.09346v2","created_at":"2026-05-18T00:32:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1709.09346","created_at":"2026-05-18T00:32:54Z"},{"alias_kind":"pith_short_12","alias_value":"MEFYDUGR2L6Z","created_at":"2026-05-18T12:31:31Z"},{"alias_kind":"pith_short_16","alias_value":"MEFYDUGR2L6ZEHUV","created_at":"2026-05-18T12:31:31Z"},{"alias_kind":"pith_short_8","alias_value":"MEFYDUGR","created_at":"2026-05-18T12:31:31Z"}],"graph_snapshots":[{"event_id":"sha256:94d76b7796d50cc3e8a2dfa9d33d634243500d513e35417d104709161195a643","target":"graph","created_at":"2026-05-18T00:32:54Z","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":"Policy-gradient approaches to reinforcement learning have two common and undesirable overhead procedures, namely warm-start training and sample variance reduction. In this paper, we describe a reinforcement learning method based on a softmax value function that requires neither of these procedures. Our method combines the advantages of policy-gradient methods with the efficiency and simplicity of maximum-likelihood approaches. We apply this new cold-start reinforcement learning method in training sequence generation models for structured output prediction problems. Empirical evidence validates","authors_text":"Nan Ding, Radu Soricut","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-09-27T06:04:03Z","title":"Cold-Start Reinforcement Learning with Softmax Policy Gradient"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1709.09346","kind":"arxiv","version":2},"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:e171c41d11d64b6fe392b488d5604fc9c5423ad66548a6251d20d1c5a27eca51","target":"record","created_at":"2026-05-18T00:32:54Z","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":"64f16bd801fa75da84b2a8469cf020371ca21730d6a2da4aca40ebb826c31f0a","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-09-27T06:04:03Z","title_canon_sha256":"68a7063ae097841732924d19cc12f11bd7d80b1650ad192ce9dd72a37b6f070a"},"schema_version":"1.0","source":{"id":"1709.09346","kind":"arxiv","version":2}},"canonical_sha256":"610b81d0d1d2fd921e9548b3e77b62390718091946bca5cfec4a1237943ec293","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"610b81d0d1d2fd921e9548b3e77b62390718091946bca5cfec4a1237943ec293","first_computed_at":"2026-05-18T00:32:54.759928Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T00:32:54.759928Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"esce+84sdx2OMBspHO5GXR9ST+NwVggDR9TVEMlZqxUTX8EabIo65eft3oXnenCL6Y0a+gQdauOvJirjiPZbBQ==","signature_status":"signed_v1","signed_at":"2026-05-18T00:32:54.760653Z","signed_message":"canonical_sha256_bytes"},"source_id":"1709.09346","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e171c41d11d64b6fe392b488d5604fc9c5423ad66548a6251d20d1c5a27eca51","sha256:94d76b7796d50cc3e8a2dfa9d33d634243500d513e35417d104709161195a643"],"state_sha256":"4289efee331ffec39ea8884d035658ea17a5b6d98f674d42a3403e49e2b298e0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"h6bSbpzdb4wyJUaUIqdPNCpTaCItI1+1Um0ThYL3daIL7mXTMJL6RXfzKhtfM8uYlKdEJaE+EYmz2pZysZJJAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-03T12:21:21.962887Z","bundle_sha256":"c9daeb58b5702d7f90f619913a30a286c7c7c77d6c63b8bb35c97a17ac3f9380"}}