{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:L3IWBGBFNRMR2ISGSYHGYN2TIH","short_pith_number":"pith:L3IWBGBF","schema_version":"1.0","canonical_sha256":"5ed16098256c591d2246960e6c375341f5c27c29fb827a47cdbf4ec499db3dae","source":{"kind":"arxiv","id":"2505.11718","version":2},"attestation_state":"computed","paper":{"title":"REMOR: Automated Peer Review Generation with LLM Reasoning and Multi-Objective Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Daniel Acuna, Pawin Taechoyotin","submitted_at":"2025-05-16T22:00:49Z","abstract_excerpt":"AI-based peer review systems tend to produce shallow and overpraising suggestions compared to human feedback. Here, we evaluate how well a reasoning LLM trained with multi-objective reinforcement learning (REMOR) can overcome these limitations. We start by designing a multi-aspect reward function that aligns with human evaluation of reviews. The aspects are related to the review itself (e.g., criticisms, novelty) and the relationship between the review and the manuscript (i.e., relevance). First, we perform supervised fine-tuning of DeepSeek-R1-Distill-Qwen-7B using LoRA on PeerRT, a new datas"},"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":"2505.11718","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-16T22:00:49Z","cross_cats_sorted":[],"title_canon_sha256":"c6ea3bc007e50c206a9e7f904ad12cf494759c0dbcaa5641954d9530ac557406","abstract_canon_sha256":"782a4291a2f99d07809ed9909f91b7a2a6cd795cfa818d578a61bb6fcc5f5e58"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:27:54.933064Z","signature_b64":"eAceBVU5CH3CFqYmZIf1z43SWnkX3jvsNS2ttTFV5r0sp5oWiW5q21YHRspTcfhh7Df19CFgYwbPP22KQNV2Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5ed16098256c591d2246960e6c375341f5c27c29fb827a47cdbf4ec499db3dae","last_reissued_at":"2026-07-05T11:27:54.932508Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:27:54.932508Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"REMOR: Automated Peer Review Generation with LLM Reasoning and Multi-Objective Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Daniel Acuna, Pawin Taechoyotin","submitted_at":"2025-05-16T22:00:49Z","abstract_excerpt":"AI-based peer review systems tend to produce shallow and overpraising suggestions compared to human feedback. Here, we evaluate how well a reasoning LLM trained with multi-objective reinforcement learning (REMOR) can overcome these limitations. We start by designing a multi-aspect reward function that aligns with human evaluation of reviews. The aspects are related to the review itself (e.g., criticisms, novelty) and the relationship between the review and the manuscript (i.e., relevance). First, we perform supervised fine-tuning of DeepSeek-R1-Distill-Qwen-7B using LoRA on PeerRT, a new datas"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.11718","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/2505.11718/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":"2505.11718","created_at":"2026-07-05T11:27:54.932563+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.11718v2","created_at":"2026-07-05T11:27:54.932563+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.11718","created_at":"2026-07-05T11:27:54.932563+00:00"},{"alias_kind":"pith_short_12","alias_value":"L3IWBGBFNRMR","created_at":"2026-07-05T11:27:54.932563+00:00"},{"alias_kind":"pith_short_16","alias_value":"L3IWBGBFNRMR2ISG","created_at":"2026-07-05T11:27:54.932563+00:00"},{"alias_kind":"pith_short_8","alias_value":"L3IWBGBF","created_at":"2026-07-05T11:27:54.932563+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25057","citing_title":"LLM-Based Scientific Peer Review: Methods, Benchmarks, and Reliability Challenges","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25415","citing_title":"LLM-as-a-Reviewer: Benchmarking Their Ability, Divergence, and Prompt Injection Resistance as Paper Reviewers","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18661","citing_title":"AI for Auto-Research: Roadmap & User Guide","ref_index":197,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24071","citing_title":"PeeriScope: A Multi-Faceted Framework for Evaluating Peer Review Quality","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH","json":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH.json","graph_json":"https://pith.science/api/pith-number/L3IWBGBFNRMR2ISGSYHGYN2TIH/graph.json","events_json":"https://pith.science/api/pith-number/L3IWBGBFNRMR2ISGSYHGYN2TIH/events.json","paper":"https://pith.science/paper/L3IWBGBF"},"agent_actions":{"view_html":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH","download_json":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH.json","view_paper":"https://pith.science/paper/L3IWBGBF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.11718&json=true","fetch_graph":"https://pith.science/api/pith-number/L3IWBGBFNRMR2ISGSYHGYN2TIH/graph.json","fetch_events":"https://pith.science/api/pith-number/L3IWBGBFNRMR2ISGSYHGYN2TIH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH/action/storage_attestation","attest_author":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH/action/author_attestation","sign_citation":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH/action/citation_signature","submit_replication":"https://pith.science/pith/L3IWBGBFNRMR2ISGSYHGYN2TIH/action/replication_record"}},"created_at":"2026-07-05T11:27:54.932563+00:00","updated_at":"2026-07-05T11:27:54.932563+00:00"}