{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:6RLUIZWFAJP2DOR6KNPZDIAT2X","short_pith_number":"pith:6RLUIZWF","schema_version":"1.0","canonical_sha256":"f4574466c5025fa1ba3e535f91a013d5d3f993144394993f0f225061af60786c","source":{"kind":"arxiv","id":"2110.02034","version":2},"attestation_state":"computed","paper":{"title":"Dropout Q-Functions for Doubly Efficient Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Taisei Hashimoto, Takahisa Imagawa, Takashi Onishi, Takuya Hiraoka, Yoshimasa Tsuruoka","submitted_at":"2021-10-05T13:28:11Z","abstract_excerpt":"Randomized ensembled double Q-learning (REDQ) (Chen et al., 2021b) has recently achieved state-of-the-art sample efficiency on continuous-action reinforcement learning benchmarks. This superior sample efficiency is made possible by using a large Q-function ensemble. However, REDQ is much less computationally efficient than non-ensemble counterparts such as Soft Actor-Critic (SAC) (Haarnoja et al., 2018a). To make REDQ more computationally efficient, we propose a method of improving computational efficiency called DroQ, which is a variant of REDQ that uses a small ensemble of dropout Q-function"},"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":"2110.02034","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-05T13:28:11Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a3dc8ba52a642581d32be79031b0f91d43587b7ba467be54f066bf1dba5f4da5","abstract_canon_sha256":"09f4c54539a828d64bd5572d34f1a484d233db132f0f2a65ff2a7abb8a73ea70"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:05:35.405196Z","signature_b64":"kualjZ51InHdTbT8kmLSJNZYsm/Ct6dmS557YbWTwr+K9ifFqyxBnaUrmLUbgj3EvrBc6olTPv8YICpe1FofAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f4574466c5025fa1ba3e535f91a013d5d3f993144394993f0f225061af60786c","last_reissued_at":"2026-07-05T04:05:35.404774Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:05:35.404774Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dropout Q-Functions for Doubly Efficient Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Taisei Hashimoto, Takahisa Imagawa, Takashi Onishi, Takuya Hiraoka, Yoshimasa Tsuruoka","submitted_at":"2021-10-05T13:28:11Z","abstract_excerpt":"Randomized ensembled double Q-learning (REDQ) (Chen et al., 2021b) has recently achieved state-of-the-art sample efficiency on continuous-action reinforcement learning benchmarks. This superior sample efficiency is made possible by using a large Q-function ensemble. However, REDQ is much less computationally efficient than non-ensemble counterparts such as Soft Actor-Critic (SAC) (Haarnoja et al., 2018a). To make REDQ more computationally efficient, we propose a method of improving computational efficiency called DroQ, which is a variant of REDQ that uses a small ensemble of dropout Q-function"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.02034","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/2110.02034/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":"2110.02034","created_at":"2026-07-05T04:05:35.404832+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.02034v2","created_at":"2026-07-05T04:05:35.404832+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.02034","created_at":"2026-07-05T04:05:35.404832+00:00"},{"alias_kind":"pith_short_12","alias_value":"6RLUIZWFAJP2","created_at":"2026-07-05T04:05:35.404832+00:00"},{"alias_kind":"pith_short_16","alias_value":"6RLUIZWFAJP2DOR6","created_at":"2026-07-05T04:05:35.404832+00:00"},{"alias_kind":"pith_short_8","alias_value":"6RLUIZWF","created_at":"2026-07-05T04:05:35.404832+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2503.15481","citing_title":"Learning to Play Piano in the Real World","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05863","citing_title":"SOPE: Stabilizing Off-Policy Evaluation for Online RL with Prior Data","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11711","citing_title":"Debiased Model-based Representations for Sample-efficient Continuous Control","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05863","citing_title":"SOPE: Stabilizing Off-Policy Evaluation for Online RL with Prior Data","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18978","citing_title":"Low-Rank Adaptation for Critic Learning in Off-Policy Reinforcement Learning","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20381","citing_title":"Distributional Value Estimation Without Target Networks for Robust Quality-Diversity","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X","json":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X.json","graph_json":"https://pith.science/api/pith-number/6RLUIZWFAJP2DOR6KNPZDIAT2X/graph.json","events_json":"https://pith.science/api/pith-number/6RLUIZWFAJP2DOR6KNPZDIAT2X/events.json","paper":"https://pith.science/paper/6RLUIZWF"},"agent_actions":{"view_html":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X","download_json":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X.json","view_paper":"https://pith.science/paper/6RLUIZWF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.02034&json=true","fetch_graph":"https://pith.science/api/pith-number/6RLUIZWFAJP2DOR6KNPZDIAT2X/graph.json","fetch_events":"https://pith.science/api/pith-number/6RLUIZWFAJP2DOR6KNPZDIAT2X/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X/action/storage_attestation","attest_author":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X/action/author_attestation","sign_citation":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X/action/citation_signature","submit_replication":"https://pith.science/pith/6RLUIZWFAJP2DOR6KNPZDIAT2X/action/replication_record"}},"created_at":"2026-07-05T04:05:35.404832+00:00","updated_at":"2026-07-05T04:05:35.404832+00:00"}