{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:O6FWNGQXZREKFDUR2DDLRBKUUD","short_pith_number":"pith:O6FWNGQX","schema_version":"1.0","canonical_sha256":"778b669a17cc48a28e91d0c6b88554a0df98c4c5700c8d4ea1fc37188ae17f78","source":{"kind":"arxiv","id":"2506.17639","version":2},"attestation_state":"computed","paper":{"title":"RLRC: Reinforcement Learning-based Recovery for Compressed Vision-Language-Action Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Xiao Li, Yixin Han, Yize Huang, Yuxuan Chen","submitted_at":"2025-06-21T08:45:32Z","abstract_excerpt":"Vision-Language-Action models (VLA) have demonstrated remarkable capabilities and strong potential in complex robotic manipulation. However, their large parameter sizes and high inference latency hinder real-world deployment, especially on resource-constrained platforms. To address this, we conduct a systematic empirical study of model compression for VLAs. Building on these insights, we present \\textit{RLRC}, a three-stage compression and recovery pipeline consisting of structured pruning, performance recovery via SFT and RL, and subsequent quantization. The RL stage incorporates a critic war"},"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":"2506.17639","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2025-06-21T08:45:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7d9c584fdc7c37f199be90d4dbd74d0cd204c74e75ae561af138fe4750c62804","abstract_canon_sha256":"b85c40299aec6e9e63e2bea0bd75c6e249e84ae1c5793208045cc2d9f5e6b321"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:09:49.734118Z","signature_b64":"j6LSiMIv1uW2W9Sr6mD+tpwEaYx+6vannX8m8Rf4jWr1ItZhQ//xm+iR7i3JN9jWs1hFR4KwhKGMXUwdAjqRAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"778b669a17cc48a28e91d0c6b88554a0df98c4c5700c8d4ea1fc37188ae17f78","last_reissued_at":"2026-06-19T16:09:49.733694Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:09:49.733694Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RLRC: Reinforcement Learning-based Recovery for Compressed Vision-Language-Action Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Xiao Li, Yixin Han, Yize Huang, Yuxuan Chen","submitted_at":"2025-06-21T08:45:32Z","abstract_excerpt":"Vision-Language-Action models (VLA) have demonstrated remarkable capabilities and strong potential in complex robotic manipulation. However, their large parameter sizes and high inference latency hinder real-world deployment, especially on resource-constrained platforms. To address this, we conduct a systematic empirical study of model compression for VLAs. Building on these insights, we present \\textit{RLRC}, a three-stage compression and recovery pipeline consisting of structured pruning, performance recovery via SFT and RL, and subsequent quantization. The RL stage incorporates a critic war"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.17639","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2506.17639/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":"2506.17639","created_at":"2026-06-19T16:09:49.733751+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.17639v2","created_at":"2026-06-19T16:09:49.733751+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.17639","created_at":"2026-06-19T16:09:49.733751+00:00"},{"alias_kind":"pith_short_12","alias_value":"O6FWNGQXZREK","created_at":"2026-06-19T16:09:49.733751+00:00"},{"alias_kind":"pith_short_16","alias_value":"O6FWNGQXZREKFDUR","created_at":"2026-06-19T16:09:49.733751+00:00"},{"alias_kind":"pith_short_8","alias_value":"O6FWNGQX","created_at":"2026-06-19T16:09:49.733751+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":6,"sample":[{"citing_arxiv_id":"2605.29662","citing_title":"SAFE-Pruner: Semantic Attention-Guided Future-Aware Token Pruning for Efficient Vision-Language-Action Manipulation","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2603.19199","citing_title":"FASTER: Rethinking Real-Time Flow VLAs","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.17486","citing_title":"DyGRO-VLA: Cross-Task Scaling of Vision-Language-Action Models via Dynamic Grouped Residual Optimization","ref_index":124,"is_internal_anchor":true},{"citing_arxiv_id":"2511.15669","citing_title":"DeepThinkVLA: Enhancing Reasoning Capability of Vision-Language-Action Models","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2603.19199","citing_title":"FASTER: Rethinking Real-Time Flow VLAs","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.01742","citing_title":"Joint Architecture-Token-Bitwidth Multi-Axis Optimization of Vision Transformers for Semiconductor IC Packaging","ref_index":19,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD","json":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD.json","graph_json":"https://pith.science/api/pith-number/O6FWNGQXZREKFDUR2DDLRBKUUD/graph.json","events_json":"https://pith.science/api/pith-number/O6FWNGQXZREKFDUR2DDLRBKUUD/events.json","paper":"https://pith.science/paper/O6FWNGQX"},"agent_actions":{"view_html":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD","download_json":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD.json","view_paper":"https://pith.science/paper/O6FWNGQX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.17639&json=true","fetch_graph":"https://pith.science/api/pith-number/O6FWNGQXZREKFDUR2DDLRBKUUD/graph.json","fetch_events":"https://pith.science/api/pith-number/O6FWNGQXZREKFDUR2DDLRBKUUD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD/action/storage_attestation","attest_author":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD/action/author_attestation","sign_citation":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD/action/citation_signature","submit_replication":"https://pith.science/pith/O6FWNGQXZREKFDUR2DDLRBKUUD/action/replication_record"}},"created_at":"2026-06-19T16:09:49.733751+00:00","updated_at":"2026-06-19T16:09:49.733751+00:00"}