{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MFAZRZIH4EOXES65YZCQX64ZAY","short_pith_number":"pith:MFAZRZIH","schema_version":"1.0","canonical_sha256":"614198e507e11d724bddc6450bfb9906376f6341dca10b7a836c392b3f8c6395","source":{"kind":"arxiv","id":"2411.17861","version":3},"attestation_state":"computed","paper":{"title":"Accelerating Proximal Policy Optimization Learning Using Task Prediction for Solving Environments with Delayed Rewards","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ahmad Ahmad, Calin Belta, Cristian-Ioan Vasile, Ho Chit Siu, Kevin Leahy, Makai Mann, Mehdi Kermanshah, Roberto Tron, Zachary Serlin","submitted_at":"2024-11-26T20:22:31Z","abstract_excerpt":"In this paper, we tackle the challenging problem of delayed rewards in reinforcement learning (RL). While Proximal Policy Optimization (PPO) has emerged as a leading Policy Gradient method, its performance can degrade under delayed rewards. We introduce two key enhancements to PPO: a hybrid policy architecture that combines an offline policy (trained on expert demonstrations) with an online PPO policy, and a reward shaping mechanism using Time Window Temporal Logic (TWTL). The hybrid architecture leverages offline data throughout training while maintaining PPO's theoretical guarantees. Buildin"},"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":"2411.17861","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-26T20:22:31Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"51ac09f0f131ce1542e3c0231447922d11e224a6c38f075de1a27ce3f57332bd","abstract_canon_sha256":"b6f018b318330b1bfac362f0fa849251230285bc22893d480ea0a436c003aa35"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:44:36.907766Z","signature_b64":"gRTyrKCO8jTcTbljiRVoNLlCSvCmo1CXuWzCPtChZ9sVaLM4PfSH56CHuDKjrcgDDryWuwUQmFKoRfnoPOJ8Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"614198e507e11d724bddc6450bfb9906376f6341dca10b7a836c392b3f8c6395","last_reissued_at":"2026-07-05T09:44:36.907260Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:44:36.907260Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Accelerating Proximal Policy Optimization Learning Using Task Prediction for Solving Environments with Delayed Rewards","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ahmad Ahmad, Calin Belta, Cristian-Ioan Vasile, Ho Chit Siu, Kevin Leahy, Makai Mann, Mehdi Kermanshah, Roberto Tron, Zachary Serlin","submitted_at":"2024-11-26T20:22:31Z","abstract_excerpt":"In this paper, we tackle the challenging problem of delayed rewards in reinforcement learning (RL). While Proximal Policy Optimization (PPO) has emerged as a leading Policy Gradient method, its performance can degrade under delayed rewards. We introduce two key enhancements to PPO: a hybrid policy architecture that combines an offline policy (trained on expert demonstrations) with an online PPO policy, and a reward shaping mechanism using Time Window Temporal Logic (TWTL). The hybrid architecture leverages offline data throughout training while maintaining PPO's theoretical guarantees. Buildin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.17861","kind":"arxiv","version":3},"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/2411.17861/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":"2411.17861","created_at":"2026-07-05T09:44:36.907328+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.17861v3","created_at":"2026-07-05T09:44:36.907328+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.17861","created_at":"2026-07-05T09:44:36.907328+00:00"},{"alias_kind":"pith_short_12","alias_value":"MFAZRZIH4EOX","created_at":"2026-07-05T09:44:36.907328+00:00"},{"alias_kind":"pith_short_16","alias_value":"MFAZRZIH4EOXES65","created_at":"2026-07-05T09:44:36.907328+00:00"},{"alias_kind":"pith_short_8","alias_value":"MFAZRZIH","created_at":"2026-07-05T09:44:36.907328+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.08561","citing_title":"ANSR-DT: A Neuro-Symbolic Framework for Adaptive and Explainable Digital Twins","ref_index":20,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY","json":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY.json","graph_json":"https://pith.science/api/pith-number/MFAZRZIH4EOXES65YZCQX64ZAY/graph.json","events_json":"https://pith.science/api/pith-number/MFAZRZIH4EOXES65YZCQX64ZAY/events.json","paper":"https://pith.science/paper/MFAZRZIH"},"agent_actions":{"view_html":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY","download_json":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY.json","view_paper":"https://pith.science/paper/MFAZRZIH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.17861&json=true","fetch_graph":"https://pith.science/api/pith-number/MFAZRZIH4EOXES65YZCQX64ZAY/graph.json","fetch_events":"https://pith.science/api/pith-number/MFAZRZIH4EOXES65YZCQX64ZAY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY/action/storage_attestation","attest_author":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY/action/author_attestation","sign_citation":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY/action/citation_signature","submit_replication":"https://pith.science/pith/MFAZRZIH4EOXES65YZCQX64ZAY/action/replication_record"}},"created_at":"2026-07-05T09:44:36.907328+00:00","updated_at":"2026-07-05T09:44:36.907328+00:00"}