{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DCPO2QGR35ACEBPYLLBHYKEI2E","short_pith_number":"pith:DCPO2QGR","schema_version":"1.0","canonical_sha256":"189eed40d1df402205f85ac27c2888d10ea5f580fa251f4144a2766abf94a1ed","source":{"kind":"arxiv","id":"2602.09023","version":4},"attestation_state":"computed","paper":{"title":"TwinRL: Digital Twin-Driven Reinforcement Learning for Real-World Robotic Manipulation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.RO","authors_text":"Chenyang Gu, Gao Huang, Jiaming Liu, Nuowei Han, Peng Jia, Qinwen Xu, Rui Zhou, Shanghang Zhang, Shaojun Shi, Shuo Gu, Sirui Han, Wenzhao Zheng, Yang Yue, Zhuoyang Liu","submitted_at":"2026-02-09T18:59:52Z","abstract_excerpt":"Despite strong generalization capabilities, Vision-Language-Action (VLA) models remain constrained by the high cost of expert demonstrations and limited real-world interaction. While online reinforcement learning (RL) has shown promise, its application to real-world VLA manipulation is hindered by low exploration efficiency and restricted exploration coverage. Through systematic real-world experiments, we observe that the effective exploration space of online RL is largely constrained by the trajectory distribution induced during supervised fine-tuning (SFT). Motivated by this observation, we "},"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":"2602.09023","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-02-09T18:59:52Z","cross_cats_sorted":[],"title_canon_sha256":"c982b7a82402dc1669fa4a21737824650a2d487cc634260587a5ec6760ba97f0","abstract_canon_sha256":"3fe65d80f961c6443c1a8d2b72c8671c3688fcd1545f68cdb2fe6dccddcf2f46"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T01:05:08.646918Z","signature_b64":"4v+liu/d35WI11d+zRN3pz/KJBPeZg1xdc50X/V0qZvCSOXWgYruttbupAPYsy5W+hvRQR9f+g0qi2HrIz4oAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"189eed40d1df402205f85ac27c2888d10ea5f580fa251f4144a2766abf94a1ed","last_reissued_at":"2026-05-20T01:05:08.645966Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T01:05:08.645966Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TwinRL: Digital Twin-Driven Reinforcement Learning for Real-World Robotic Manipulation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.RO","authors_text":"Chenyang Gu, Gao Huang, Jiaming Liu, Nuowei Han, Peng Jia, Qinwen Xu, Rui Zhou, Shanghang Zhang, Shaojun Shi, Shuo Gu, Sirui Han, Wenzhao Zheng, Yang Yue, Zhuoyang Liu","submitted_at":"2026-02-09T18:59:52Z","abstract_excerpt":"Despite strong generalization capabilities, Vision-Language-Action (VLA) models remain constrained by the high cost of expert demonstrations and limited real-world interaction. While online reinforcement learning (RL) has shown promise, its application to real-world VLA manipulation is hindered by low exploration efficiency and restricted exploration coverage. Through systematic real-world experiments, we observe that the effective exploration space of online RL is largely constrained by the trajectory distribution induced during supervised fine-tuning (SFT). Motivated by this observation, we "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.09023","kind":"arxiv","version":4},"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/2602.09023/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2602.09023","created_at":"2026-05-20T01:05:08.646086+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.09023v4","created_at":"2026-05-20T01:05:08.646086+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.09023","created_at":"2026-05-20T01:05:08.646086+00:00"},{"alias_kind":"pith_short_12","alias_value":"DCPO2QGR35AC","created_at":"2026-05-20T01:05:08.646086+00:00"},{"alias_kind":"pith_short_16","alias_value":"DCPO2QGR35ACEBPY","created_at":"2026-05-20T01:05:08.646086+00:00"},{"alias_kind":"pith_short_8","alias_value":"DCPO2QGR","created_at":"2026-05-20T01:05:08.646086+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2604.28192","citing_title":"LaST-R1: Reinforcing Robotic Manipulation via Adaptive Physical Latent Reasoning","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2604.28192","citing_title":"LaST-R1: Reinforcing Robotic Manipulation via Adaptive Physical Latent Reasoning","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2604.10165","citing_title":"MoRI: Mixture of RL and IL Experts for Long-Horizon Manipulation Tasks","ref_index":7,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E","json":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E.json","graph_json":"https://pith.science/api/pith-number/DCPO2QGR35ACEBPYLLBHYKEI2E/graph.json","events_json":"https://pith.science/api/pith-number/DCPO2QGR35ACEBPYLLBHYKEI2E/events.json","paper":"https://pith.science/paper/DCPO2QGR"},"agent_actions":{"view_html":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E","download_json":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E.json","view_paper":"https://pith.science/paper/DCPO2QGR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.09023&json=true","fetch_graph":"https://pith.science/api/pith-number/DCPO2QGR35ACEBPYLLBHYKEI2E/graph.json","fetch_events":"https://pith.science/api/pith-number/DCPO2QGR35ACEBPYLLBHYKEI2E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E/action/storage_attestation","attest_author":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E/action/author_attestation","sign_citation":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E/action/citation_signature","submit_replication":"https://pith.science/pith/DCPO2QGR35ACEBPYLLBHYKEI2E/action/replication_record"}},"created_at":"2026-05-20T01:05:08.646086+00:00","updated_at":"2026-05-20T01:05:08.646086+00:00"}