{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6EQULQPSREPGRMBSQR2EEDJPLN","short_pith_number":"pith:6EQULQPS","schema_version":"1.0","canonical_sha256":"f12145c1f2891e68b0328474420d2f5b421bbde7d9847e8181cfb93a1b1b1014","source":{"kind":"arxiv","id":"2607.09042","version":1},"attestation_state":"computed","paper":{"title":"Learning More from Less: Reinforcement Learning from Hindsight","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alex Pentland, Iris Xu, John Marangola, Nitish Dashora, Pulkit Agrawal, Richard Li, Sunshine Jiang, Thomas Liu, Yuheng Zhi, Zexue He, Zhang-Wei Hong","submitted_at":"2026-07-10T02:17:41Z","abstract_excerpt":"Reinforcement learning (RL) is increasingly used to post-train vision-language-action (VLA) models, but every update consumes robot rollouts that are slow and costly to collect, making sample efficiency a central concern. Manipulation tasks typically provide only sparse rewards, so a weak policy fails almost every rollout early in training and has little to learn from, even when those failures execute coherent behavior. Such a failure, however, is a success at a different task. We present Learning from Hindsight (LfH), which brings hindsight relabeling to RL post-training of VLAs by scoring fa"},"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":"2607.09042","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-10T02:17:41Z","cross_cats_sorted":[],"title_canon_sha256":"6eb32790b5dc73f111f29a2cbb646af083098fcc3883126558b3b27f9782088e","abstract_canon_sha256":"a63335393536bf6670bc3acf1046c5f000f0921f07f2f7acfe2cffa1994b03da"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-13T00:17:37.144632Z","signature_b64":"yNtTMU/W0RpZ3pLxzA2/pqfzeSUJ3Ihb1C9rS4fXsCiZDXdi2dcTaD3SAusb8Ff9LmO9wONp31IvDPBRtSRXAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f12145c1f2891e68b0328474420d2f5b421bbde7d9847e8181cfb93a1b1b1014","last_reissued_at":"2026-07-13T00:17:37.143391Z","signature_status":"signed_v1","first_computed_at":"2026-07-13T00:17:37.143391Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning More from Less: Reinforcement Learning from Hindsight","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alex Pentland, Iris Xu, John Marangola, Nitish Dashora, Pulkit Agrawal, Richard Li, Sunshine Jiang, Thomas Liu, Yuheng Zhi, Zexue He, Zhang-Wei Hong","submitted_at":"2026-07-10T02:17:41Z","abstract_excerpt":"Reinforcement learning (RL) is increasingly used to post-train vision-language-action (VLA) models, but every update consumes robot rollouts that are slow and costly to collect, making sample efficiency a central concern. Manipulation tasks typically provide only sparse rewards, so a weak policy fails almost every rollout early in training and has little to learn from, even when those failures execute coherent behavior. Such a failure, however, is a success at a different task. We present Learning from Hindsight (LfH), which brings hindsight relabeling to RL post-training of VLAs by scoring fa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.09042","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/2607.09042/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":"2607.09042","created_at":"2026-07-13T00:17:37.144031+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.09042v1","created_at":"2026-07-13T00:17:37.144031+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.09042","created_at":"2026-07-13T00:17:37.144031+00:00"},{"alias_kind":"pith_short_12","alias_value":"6EQULQPSREPG","created_at":"2026-07-13T00:17:37.144031+00:00"},{"alias_kind":"pith_short_16","alias_value":"6EQULQPSREPGRMBS","created_at":"2026-07-13T00:17:37.144031+00:00"},{"alias_kind":"pith_short_8","alias_value":"6EQULQPS","created_at":"2026-07-13T00:17:37.144031+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN","json":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN.json","graph_json":"https://pith.science/api/pith-number/6EQULQPSREPGRMBSQR2EEDJPLN/graph.json","events_json":"https://pith.science/api/pith-number/6EQULQPSREPGRMBSQR2EEDJPLN/events.json","paper":"https://pith.science/paper/6EQULQPS"},"agent_actions":{"view_html":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN","download_json":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN.json","view_paper":"https://pith.science/paper/6EQULQPS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.09042&json=true","fetch_graph":"https://pith.science/api/pith-number/6EQULQPSREPGRMBSQR2EEDJPLN/graph.json","fetch_events":"https://pith.science/api/pith-number/6EQULQPSREPGRMBSQR2EEDJPLN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN/action/storage_attestation","attest_author":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN/action/author_attestation","sign_citation":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN/action/citation_signature","submit_replication":"https://pith.science/pith/6EQULQPSREPGRMBSQR2EEDJPLN/action/replication_record"}},"created_at":"2026-07-13T00:17:37.144031+00:00","updated_at":"2026-07-13T00:17:37.144031+00:00"}