{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:2N2DMZRLT7ZINZ5NB2K37CIEJ5","short_pith_number":"pith:2N2DMZRL","schema_version":"1.0","canonical_sha256":"d37436662b9ff286e7ad0e95bf89044f6d335927bf43e0e646358cac5e62d640","source":{"kind":"arxiv","id":"2506.08737","version":1},"attestation_state":"computed","paper":{"title":"Exploration by Random Reward Perturbation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Guoji Fu, Haozhe Ma, Jiele Wu, Tze-Yun Leong, Zhengding Luo","submitted_at":"2025-06-10T12:34:00Z","abstract_excerpt":"We introduce Random Reward Perturbation (RRP), a novel exploration strategy for reinforcement learning (RL). Our theoretical analyses demonstrate that adding zero-mean noise to environmental rewards effectively enhances policy diversity during training, thereby expanding the range of exploration. RRP is fully compatible with the action-perturbation-based exploration strategies, such as $\\epsilon$-greedy, stochastic policies, and entropy regularization, providing additive improvements to exploration effects. It is general, lightweight, and can be integrated into existing RL algorithms with mini"},"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":"2506.08737","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-10T12:34:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c80b53668d1db31b030ecba6cfb8314de1862a3bafe52a5e65299b79746e4428","abstract_canon_sha256":"f6ed7d4c1932cb4519151401a8c706619c470f8f889a7ce13fb202f35665c4c4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:19:12.447692Z","signature_b64":"dWUdGlJj+U66K/SsCjUqupWwkXkata5H3/FWSoO8EZD/rRaGHjAhHvOgSRn2mJEdM8juK8DA7ePgeqV1RAdfCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d37436662b9ff286e7ad0e95bf89044f6d335927bf43e0e646358cac5e62d640","last_reissued_at":"2026-07-05T11:19:12.447220Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:19:12.447220Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Exploration by Random Reward Perturbation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Guoji Fu, Haozhe Ma, Jiele Wu, Tze-Yun Leong, Zhengding Luo","submitted_at":"2025-06-10T12:34:00Z","abstract_excerpt":"We introduce Random Reward Perturbation (RRP), a novel exploration strategy for reinforcement learning (RL). Our theoretical analyses demonstrate that adding zero-mean noise to environmental rewards effectively enhances policy diversity during training, thereby expanding the range of exploration. RRP is fully compatible with the action-perturbation-based exploration strategies, such as $\\epsilon$-greedy, stochastic policies, and entropy regularization, providing additive improvements to exploration effects. It is general, lightweight, and can be integrated into existing RL algorithms with mini"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.08737","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/2506.08737/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":"2506.08737","created_at":"2026-07-05T11:19:12.447273+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.08737v1","created_at":"2026-07-05T11:19:12.447273+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.08737","created_at":"2026-07-05T11:19:12.447273+00:00"},{"alias_kind":"pith_short_12","alias_value":"2N2DMZRLT7ZI","created_at":"2026-07-05T11:19:12.447273+00:00"},{"alias_kind":"pith_short_16","alias_value":"2N2DMZRLT7ZINZ5N","created_at":"2026-07-05T11:19:12.447273+00:00"},{"alias_kind":"pith_short_8","alias_value":"2N2DMZRL","created_at":"2026-07-05T11:19:12.447273+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.04175","citing_title":"AD-FM: Multimodal LLMs for Anomaly Detection via Multi-Stage Reasoning and Fine-Grained Reward Optimization","ref_index":33,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5","json":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5.json","graph_json":"https://pith.science/api/pith-number/2N2DMZRLT7ZINZ5NB2K37CIEJ5/graph.json","events_json":"https://pith.science/api/pith-number/2N2DMZRLT7ZINZ5NB2K37CIEJ5/events.json","paper":"https://pith.science/paper/2N2DMZRL"},"agent_actions":{"view_html":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5","download_json":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5.json","view_paper":"https://pith.science/paper/2N2DMZRL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.08737&json=true","fetch_graph":"https://pith.science/api/pith-number/2N2DMZRLT7ZINZ5NB2K37CIEJ5/graph.json","fetch_events":"https://pith.science/api/pith-number/2N2DMZRLT7ZINZ5NB2K37CIEJ5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5/action/storage_attestation","attest_author":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5/action/author_attestation","sign_citation":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5/action/citation_signature","submit_replication":"https://pith.science/pith/2N2DMZRLT7ZINZ5NB2K37CIEJ5/action/replication_record"}},"created_at":"2026-07-05T11:19:12.447273+00:00","updated_at":"2026-07-05T11:19:12.447273+00:00"}