{"doi":"10.1609/aaai.v34i05.6249","title":"Sequence Generation with Optimal-Transport-Enhanced Reinforcement Learning","abstract":"Reinforcement learning (RL) has been widely used to aid training in language generation. This is achieved by enhancing standard maximum likelihood objectives with user-specified reward functions that encourage global semantic consistency. We propose a principled approach to address the difficulties associated with RL-based solutions, namely, high-variance gradients, uninformative rewards and brittle training. By leveraging the optimal transport distance, we introduce a regularizer that significantly alleviates the above issues. Our formulation emphasizes the preservation of semantic features, enabling end-to-end training instead of ad-hoc fine-tuning, and when combined with RL, it controls the exploration space for more efficient model updates. To validate the effectiveness of the proposed solution, we perform a comprehensive evaluation covering a wide variety of NLP tasks: machine translation, abstractive text summarization and image caption, with consistent improvements over competing solutions.","journal":"Proceedings of the AAAI Conference on Artificial Intelligence","year":2020,"id":119971,"datarank":0.3958585994422889,"base_score":2.639057329615259,"endowment":2.639057329615259,"self_citation_contribution":0.3958585994422889,"citation_network_contribution":0.0,"self_endowment_contribution":0.3958585994422889,"citer_contribution":0.0,"corpus_percentile":null,"corpus_rank":null,"citation_count":13,"citer_count":0,"citers_with_citation_signal":0,"citers_with_endowment":0,"datacite_reuse_total":0,"is_dataset":false,"is_dataset_confidence":0.948,"is_data_producer":false,"deposit_databanks":null,"is_oa":true,"file_count":0,"downloads":0,"has_version_chain":false,"published_date":"2020-01-01","fair_score":null,"fair_percentile":null,"algorithm_id":"datarank_citation_only_1hop_v6","ranking_scope":"data_only","authors":[{"id":556814,"name":"Ke Bai","orcid":null,"position":1,"is_corresponding":false},{"id":556234,"name":"Chenyang Tao","orcid":"0000-0002-2155-8180","position":2,"is_corresponding":false},{"id":551733,"name":"Yizhe Zhang","orcid":"0000-0002-9599-7995","position":3,"is_corresponding":false},{"id":555455,"name":"Guoyin Wang","orcid":"0000-0002-8521-5232","position":4,"is_corresponding":false},{"id":556235,"name":"Wenlin Wang","orcid":"0000-0003-0838-1920","position":5,"is_corresponding":false},{"id":331976,"name":"Ricardo Henao","orcid":"0000-0003-4980-845X","position":6,"is_corresponding":false},{"id":513735,"name":"Lawrence Carin","orcid":"0000-0001-6277-7948","position":7,"is_corresponding":false},{"id":556233,"name":"Li‐Qun Chen","orcid":"0000-0002-3694-0833","position":0,"is_corresponding":true}],"reference_count":116,"raw_metadata":{"citation_network_status":"fetched"},"created_at":"2026-07-18T23:14:13.002105Z","pmid":null,"pmcid":null,"fwci":null,"citation_percentile":null,"influential_citations":0,"oa_status":null,"license":null,"views":0,"total_file_size_bytes":0,"version_count":0,"fair_f":null,"fair_a":null,"fair_i":null,"fair_r":null,"fair_zscore":null,"fair_rationale":null,"fair_model":null,"fair_agent_version":null,"fair_fulltext_source":null,"fair_has_llm":null,"fair_computed_at":null,"clinical_trials":[],"software_tools":[],"db_accessions":[],"linked_datasets":[],"topics":[]}