{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:USLONIBFBKNHJZFHPPVXGZG7BR","short_pith_number":"pith:USLONIBF","canonical_record":{"source":{"id":"2606.01151","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-31T10:40:28Z","cross_cats_sorted":[],"title_canon_sha256":"9f260fa9c28358becc92f6ee1ef36bd4ec09d93e8e7d6f0185963469cf47bb42","abstract_canon_sha256":"b1104f07489ac1d85e257638c4c9019be45fbfebae7680b65baedb06c1025d5d"},"schema_version":"1.0"},"canonical_sha256":"a496e6a0250a9a74e4a77beb7364df0c7a382205975917d4fe1ef7291e50b9f8","source":{"kind":"arxiv","id":"2606.01151","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.01151","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"arxiv_version","alias_value":"2606.01151v1","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.01151","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"pith_short_12","alias_value":"USLONIBFBKNH","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"pith_short_16","alias_value":"USLONIBFBKNHJZFH","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"pith_short_8","alias_value":"USLONIBF","created_at":"2026-06-02T02:04:25Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:USLONIBFBKNHJZFHPPVXGZG7BR","target":"record","payload":{"canonical_record":{"source":{"id":"2606.01151","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-31T10:40:28Z","cross_cats_sorted":[],"title_canon_sha256":"9f260fa9c28358becc92f6ee1ef36bd4ec09d93e8e7d6f0185963469cf47bb42","abstract_canon_sha256":"b1104f07489ac1d85e257638c4c9019be45fbfebae7680b65baedb06c1025d5d"},"schema_version":"1.0"},"canonical_sha256":"a496e6a0250a9a74e4a77beb7364df0c7a382205975917d4fe1ef7291e50b9f8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:25.153301Z","signature_b64":"CW39NssjqJQxnRmok0gbX5tij8ohik6zk5V2Gr0AcM+gqsLIU6LpTsBLsgz+L/greYsy5TGMWgqBWokRAg4FDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a496e6a0250a9a74e4a77beb7364df0c7a382205975917d4fe1ef7291e50b9f8","last_reissued_at":"2026-06-02T02:04:25.152885Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:25.152885Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.01151","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":"2026-06-02T02:04:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ANQ+pIM7pNOMrQG+wqtbzJI76i0r/Hos7XHE7ZJCCOvAkHUuydrNZM0a+CA158rR+R6Wb6QPyP0Utl5hX3Q7Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-30T17:10:54.272290Z"},"content_sha256":"6e4aa05e9929df304109603faeb3fe93da66e2d7365d03e4781bcfdb2c8a4cfe","schema_version":"1.0","event_id":"sha256:6e4aa05e9929df304109603faeb3fe93da66e2d7365d03e4781bcfdb2c8a4cfe"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:USLONIBFBKNHJZFHPPVXGZG7BR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Lagrangian Perturbation Diffusion Steering: Latent Reinforcement Learning for Generative Policies","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hikmet Simsir, Ozgur S. Oguz","submitted_at":"2026-05-31T10:40:28Z","abstract_excerpt":"Behavior cloning with high-capacity generative policies achieves strong imitation performance, but is often limited by demonstration coverage and distribution shift. Direct reinforcement learning fine-tuning can improve performance, but updating large action decoders is frequently unstable and sample inefficient. We propose Lagrangian Perturbation Diffusion Steering (LP-DS), a lightweight adaptation method that improves a frozen generative policy by learning a compact noise-space perturbation before decoding. LP-DS optimizes this perturbation with a Lagrangian trust-region objective, improving"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.01151","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/2606.01151/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":"2026-06-02T02:04:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FQTuRu3jydX+S0MOsq96NDMAXCa2fHXCArkounL6l/qwklgsYanrYFbuR93CqGnzHEvy9S31Efncyxubd2uvBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-30T17:10:54.272678Z"},"content_sha256":"5bff3de1840a8b3f40a91da1ee49b5f9b3270ef3371d01514210ee9eee95ac25","schema_version":"1.0","event_id":"sha256:5bff3de1840a8b3f40a91da1ee49b5f9b3270ef3371d01514210ee9eee95ac25"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/USLONIBFBKNHJZFHPPVXGZG7BR/bundle.json","state_url":"https://pith.science/pith/USLONIBFBKNHJZFHPPVXGZG7BR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/USLONIBFBKNHJZFHPPVXGZG7BR/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":"2026-06-30T17:10:54Z","links":{"resolver":"https://pith.science/pith/USLONIBFBKNHJZFHPPVXGZG7BR","bundle":"https://pith.science/pith/USLONIBFBKNHJZFHPPVXGZG7BR/bundle.json","state":"https://pith.science/pith/USLONIBFBKNHJZFHPPVXGZG7BR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/USLONIBFBKNHJZFHPPVXGZG7BR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:USLONIBFBKNHJZFHPPVXGZG7BR","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":"b1104f07489ac1d85e257638c4c9019be45fbfebae7680b65baedb06c1025d5d","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-31T10:40:28Z","title_canon_sha256":"9f260fa9c28358becc92f6ee1ef36bd4ec09d93e8e7d6f0185963469cf47bb42"},"schema_version":"1.0","source":{"id":"2606.01151","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.01151","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"arxiv_version","alias_value":"2606.01151v1","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.01151","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"pith_short_12","alias_value":"USLONIBFBKNH","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"pith_short_16","alias_value":"USLONIBFBKNHJZFH","created_at":"2026-06-02T02:04:25Z"},{"alias_kind":"pith_short_8","alias_value":"USLONIBF","created_at":"2026-06-02T02:04:25Z"}],"graph_snapshots":[{"event_id":"sha256:5bff3de1840a8b3f40a91da1ee49b5f9b3270ef3371d01514210ee9eee95ac25","target":"graph","created_at":"2026-06-02T02:04:25Z","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/2606.01151/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Behavior cloning with high-capacity generative policies achieves strong imitation performance, but is often limited by demonstration coverage and distribution shift. Direct reinforcement learning fine-tuning can improve performance, but updating large action decoders is frequently unstable and sample inefficient. We propose Lagrangian Perturbation Diffusion Steering (LP-DS), a lightweight adaptation method that improves a frozen generative policy by learning a compact noise-space perturbation before decoding. LP-DS optimizes this perturbation with a Lagrangian trust-region objective, improving","authors_text":"Hikmet Simsir, Ozgur S. Oguz","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-31T10:40:28Z","title":"Lagrangian Perturbation Diffusion Steering: Latent Reinforcement Learning for Generative Policies"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.01151","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":"sha256:6e4aa05e9929df304109603faeb3fe93da66e2d7365d03e4781bcfdb2c8a4cfe","target":"record","created_at":"2026-06-02T02:04:25Z","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":"b1104f07489ac1d85e257638c4c9019be45fbfebae7680b65baedb06c1025d5d","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-31T10:40:28Z","title_canon_sha256":"9f260fa9c28358becc92f6ee1ef36bd4ec09d93e8e7d6f0185963469cf47bb42"},"schema_version":"1.0","source":{"id":"2606.01151","kind":"arxiv","version":1}},"canonical_sha256":"a496e6a0250a9a74e4a77beb7364df0c7a382205975917d4fe1ef7291e50b9f8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a496e6a0250a9a74e4a77beb7364df0c7a382205975917d4fe1ef7291e50b9f8","first_computed_at":"2026-06-02T02:04:25.152885Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T02:04:25.152885Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"CW39NssjqJQxnRmok0gbX5tij8ohik6zk5V2Gr0AcM+gqsLIU6LpTsBLsgz+L/greYsy5TGMWgqBWokRAg4FDg==","signature_status":"signed_v1","signed_at":"2026-06-02T02:04:25.153301Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.01151","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6e4aa05e9929df304109603faeb3fe93da66e2d7365d03e4781bcfdb2c8a4cfe","sha256:5bff3de1840a8b3f40a91da1ee49b5f9b3270ef3371d01514210ee9eee95ac25"],"state_sha256":"285d88ca7e11f5a4da05f9179fa8393fdd17977e9659cef14c2d88e96913632a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tYXNY6RvOjbycAjmZx1ansVT2hIY9upgPKzH8iI35u8Qd15nG1DXU35FUONnWDpZLENTFCEnfXoWdZOAEJJcDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-30T17:10:54.274699Z","bundle_sha256":"084e9f25b7803603b55fa931ed9ff05f968450bce0d7ff70fafcba70c25bdf29"}}