{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:KH3JMQ5LKPD7IU2645ZASETSVE","short_pith_number":"pith:KH3JMQ5L","canonical_record":{"source":{"id":"2605.30154","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T16:14:45Z","cross_cats_sorted":[],"title_canon_sha256":"4ad11462a8d826374e56ad7c6850b30218a158526e86ee1e00d447947deaf924","abstract_canon_sha256":"700fd5505fb1825b7f6c60be449d554fd63a66df753c2aa8560e5fe941785edb"},"schema_version":"1.0"},"canonical_sha256":"51f69643ab53c7f4535ee772091272a93e21c04ae7854e5efa84e4d3c70bca12","source":{"kind":"arxiv","id":"2605.30154","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.30154","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"arxiv_version","alias_value":"2605.30154v1","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30154","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"pith_short_12","alias_value":"KH3JMQ5LKPD7","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"pith_short_16","alias_value":"KH3JMQ5LKPD7IU26","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"pith_short_8","alias_value":"KH3JMQ5L","created_at":"2026-05-29T02:06:11Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:KH3JMQ5LKPD7IU2645ZASETSVE","target":"record","payload":{"canonical_record":{"source":{"id":"2605.30154","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T16:14:45Z","cross_cats_sorted":[],"title_canon_sha256":"4ad11462a8d826374e56ad7c6850b30218a158526e86ee1e00d447947deaf924","abstract_canon_sha256":"700fd5505fb1825b7f6c60be449d554fd63a66df753c2aa8560e5fe941785edb"},"schema_version":"1.0"},"canonical_sha256":"51f69643ab53c7f4535ee772091272a93e21c04ae7854e5efa84e4d3c70bca12","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T02:06:11.238726Z","signature_b64":"hpXTuEkJyZGLtO8G2Oaj+IngVsNHlsjlbyw4mHwrT30McposGi7tmr7u2CuRwnNtVyiySZ98aVbrOtyqYt1eDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"51f69643ab53c7f4535ee772091272a93e21c04ae7854e5efa84e4d3c70bca12","last_reissued_at":"2026-05-29T02:06:11.238323Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T02:06:11.238323Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.30154","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-29T02:06:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"EfjLAfwcw+nmLOVhQeZC/2b1bIKhjjhD+QNWDl4lIZcCSDT2hneutsDzyMet/G0r3glrMBV05uEbJao2O8noDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-29T20:38:22.360592Z"},"content_sha256":"ec5813b258d9cd693ce1c9bbf67b5e4d026863443c40d982b13df311d56c0dba","schema_version":"1.0","event_id":"sha256:ec5813b258d9cd693ce1c9bbf67b5e4d026863443c40d982b13df311d56c0dba"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:KH3JMQ5LKPD7IU2645ZASETSVE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RL2ML: Finite-Rollout Surrogate Objectives from Reinforcement Learning to Maximum Likelihood","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Yifu Zheng","submitted_at":"2026-05-28T16:14:45Z","abstract_excerpt":"Correctness-based Reinforcement Learning with Verifiable Rewards (RLVR) trains language models from binary feedback on sampled outputs, but the objective optimized in expectation and the stochastic update geometry induced by finite rollout groups are often conflated. This paper develops RL2ML, a family of finite-rollout surrogate objectives with a closed-form, exactly unbiased gradient estimator. The family continuously connects standard reinforcement learning, maximum-likelihood-like training, and beyond-maximum-likelihood objectives while preserving estimator-objective alignment under a fixe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30154","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/2605.30154/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-29T02:06:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8H1SSnvc6jzyktWQJ/N/ivY687V/YZTD6CB/s1ZIPVkwCcT6KZrBAUxzH7QgoihS/NhmTynesNr/Gc0dyieTAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-29T20:38:22.360972Z"},"content_sha256":"2fba3987f126884700e4d60622b6cf7949eb9b4c5ddee16176eb0e34eb7e9b00","schema_version":"1.0","event_id":"sha256:2fba3987f126884700e4d60622b6cf7949eb9b4c5ddee16176eb0e34eb7e9b00"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KH3JMQ5LKPD7IU2645ZASETSVE/bundle.json","state_url":"https://pith.science/pith/KH3JMQ5LKPD7IU2645ZASETSVE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KH3JMQ5LKPD7IU2645ZASETSVE/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":"2026-06-29T20:38:22Z","links":{"resolver":"https://pith.science/pith/KH3JMQ5LKPD7IU2645ZASETSVE","bundle":"https://pith.science/pith/KH3JMQ5LKPD7IU2645ZASETSVE/bundle.json","state":"https://pith.science/pith/KH3JMQ5LKPD7IU2645ZASETSVE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KH3JMQ5LKPD7IU2645ZASETSVE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:KH3JMQ5LKPD7IU2645ZASETSVE","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":"700fd5505fb1825b7f6c60be449d554fd63a66df753c2aa8560e5fe941785edb","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T16:14:45Z","title_canon_sha256":"4ad11462a8d826374e56ad7c6850b30218a158526e86ee1e00d447947deaf924"},"schema_version":"1.0","source":{"id":"2605.30154","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.30154","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"arxiv_version","alias_value":"2605.30154v1","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30154","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"pith_short_12","alias_value":"KH3JMQ5LKPD7","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"pith_short_16","alias_value":"KH3JMQ5LKPD7IU26","created_at":"2026-05-29T02:06:11Z"},{"alias_kind":"pith_short_8","alias_value":"KH3JMQ5L","created_at":"2026-05-29T02:06:11Z"}],"graph_snapshots":[{"event_id":"sha256:2fba3987f126884700e4d60622b6cf7949eb9b4c5ddee16176eb0e34eb7e9b00","target":"graph","created_at":"2026-05-29T02:06:11Z","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"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2605.30154/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Correctness-based Reinforcement Learning with Verifiable Rewards (RLVR) trains language models from binary feedback on sampled outputs, but the objective optimized in expectation and the stochastic update geometry induced by finite rollout groups are often conflated. This paper develops RL2ML, a family of finite-rollout surrogate objectives with a closed-form, exactly unbiased gradient estimator. The family continuously connects standard reinforcement learning, maximum-likelihood-like training, and beyond-maximum-likelihood objectives while preserving estimator-objective alignment under a fixe","authors_text":"Yifu Zheng","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T16:14:45Z","title":"RL2ML: Finite-Rollout Surrogate Objectives from Reinforcement Learning to Maximum Likelihood"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30154","kind":"arxiv","version":1},"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:ec5813b258d9cd693ce1c9bbf67b5e4d026863443c40d982b13df311d56c0dba","target":"record","created_at":"2026-05-29T02:06:11Z","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":"700fd5505fb1825b7f6c60be449d554fd63a66df753c2aa8560e5fe941785edb","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-28T16:14:45Z","title_canon_sha256":"4ad11462a8d826374e56ad7c6850b30218a158526e86ee1e00d447947deaf924"},"schema_version":"1.0","source":{"id":"2605.30154","kind":"arxiv","version":1}},"canonical_sha256":"51f69643ab53c7f4535ee772091272a93e21c04ae7854e5efa84e4d3c70bca12","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"51f69643ab53c7f4535ee772091272a93e21c04ae7854e5efa84e4d3c70bca12","first_computed_at":"2026-05-29T02:06:11.238323Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-29T02:06:11.238323Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"hpXTuEkJyZGLtO8G2Oaj+IngVsNHlsjlbyw4mHwrT30McposGi7tmr7u2CuRwnNtVyiySZ98aVbrOtyqYt1eDQ==","signature_status":"signed_v1","signed_at":"2026-05-29T02:06:11.238726Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.30154","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ec5813b258d9cd693ce1c9bbf67b5e4d026863443c40d982b13df311d56c0dba","sha256:2fba3987f126884700e4d60622b6cf7949eb9b4c5ddee16176eb0e34eb7e9b00"],"state_sha256":"83e90178482c203f008b13d0d9e2b4302075a81431939eeae30292e301a0d180"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pZ4aHDjXornUYwTpr+JMGritJhcK6aOun6iVXqRUa74cOo4QmYBi7hxSjygltvlsIi3xL7NGnjLU1f1KUIrUBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-29T20:38:22.363047Z","bundle_sha256":"64274bce375c04adb576e505ebe6a83418c0ad3431ba118bf585b118fb2ce826"}}