{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,9]],"date-time":"2026-04-09T14:38:13Z","timestamp":1775745493147,"version":"3.50.1"},"reference-count":50,"publisher":"MDPI AG","issue":"12","license":[{"start":{"date-parts":[[2025,12,15]],"date-time":"2025-12-15T00:00:00Z","timestamp":1765756800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Algorithms"],"abstract":"<jats:p>Effective financial fraud detection requires systems that can interpret complex transaction semantics while dynamically adapting to asymmetric operational costs. We propose a hybrid framework in which a large language model (LLM) serves as an encoder, transforming heterogeneous transaction data into a unified embedding space. These embeddings define the state representation for a reinforcement learning (RL) agent, which acts as a fraud classifier optimized with business-aligned rewards that heavily penalize false negatives while controlling false positives. We evaluate the approach on two benchmark datasets\u2014European Credit Card Fraud and PaySim\u2014demonstrating that policy-gradient methods, particularly A2C, achieve high recall without sacrificing precision. Critically, our ablation study reveals that this hybrid architecture yields substantial performance gains on semantically rich transaction logs, whereas the advantage diminishes on mathematically compressed, anonymized features. Our results highlight the potential of coupling LLM-driven representations with RL policies for cost-sensitive and adaptive fraud detection.<\/jats:p>","DOI":"10.3390\/a18120792","type":"journal-article","created":{"date-parts":[[2025,12,16]],"date-time":"2025-12-16T08:46:52Z","timestamp":1765874812000},"page":"792","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["LLM-Assisted Financial Fraud Detection with Reinforcement Learning"],"prefix":"10.3390","volume":"18","author":[{"ORCID":"https:\/\/orcid.org\/0009-0007-1243-6281","authenticated-orcid":false,"given":"Ahmed Djalal","family":"Hacini","sequence":"first","affiliation":[{"name":"Intelligent Systems Engineering Department, National Higher School of Artificial Intelligence (ENSIA), Algiers 16093, Algeria"}]},{"given":"Mohamed","family":"Benabdelouahad","sequence":"additional","affiliation":[{"name":"Intelligent Systems Engineering Department, National Higher School of Artificial Intelligence (ENSIA), Algiers 16093, Algeria"}]},{"given":"Ishak","family":"Abassi","sequence":"additional","affiliation":[{"name":"Intelligent Systems Engineering Department, National Higher School of Artificial Intelligence (ENSIA), Algiers 16093, Algeria"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-9390-3282","authenticated-orcid":false,"given":"Sohaib","family":"Houhou","sequence":"additional","affiliation":[{"name":"Intelligent Systems Engineering Department, National Higher School of Artificial Intelligence (ENSIA), Algiers 16093, Algeria"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4920-1850","authenticated-orcid":false,"given":"Aissa","family":"Boulmerka","sequence":"additional","affiliation":[{"name":"Intelligent Systems Engineering Department, National Higher School of Artificial Intelligence (ENSIA), Algiers 16093, Algeria"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0309-8942","authenticated-orcid":false,"given":"Nadir","family":"Farhi","sequence":"additional","affiliation":[{"name":"Cosys-Grettia, University Gustave Eiffel, F-77454 Marne-la-Vall\u00e9e, France"}]}],"member":"1968","published-online":{"date-parts":[[2025,12,15]]},"reference":[{"key":"ref_1","unstructured":"Dal Pozzolo, A., Johnson, R., Caelen, O., and Bontempi, G. (2025, April 01). Credit Card Fraud Detection Dataset, 2015. Available online: https:\/\/www.kaggle.com\/datasets\/mlg-ulb\/creditcardfraud."},{"key":"ref_2","unstructured":"Lopez-Rojas, E.A., Elmir, A., and Axelsson, S. (2016, January 26\u201328). PaySim: A financial mobile money simulator for fraud detection. Proceedings of the European Modeling and Simulation Symposium (EMSS), Larnaca, Cyprus."},{"key":"ref_3","unstructured":"ACI Worldwide (2025, April 01). Strategies for Fighting Fraud in the Real-Time World; Technical Report. Available online: https:\/\/www.aciworldwide.com\/wp-content\/uploads\/2021\/04\/strategies-for-fighting-fraud-in-a-real-time-world.pdf."},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Popat, R.R., and Chaudhary, J. (2018, January 11\u201312). A Survey on Credit Card Fraud Detection Using Machine Learning. Proceedings of the 2018 2nd International Conference on Trends in Electronics and Informatics (ICOEI), Tirunelveli, India.","DOI":"10.1109\/ICOEI.2018.8553963"},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"937","DOI":"10.1007\/s13198-016-0551-y","article-title":"A survey of machine-learning and nature-inspired based credit card fraud detection techniques","volume":"8","author":"Adewumi","year":"2017","journal-title":"Int. J. Syst. Assur. Eng. Manag."},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Rojan, Z. (2024). Financial Fraud Detection Based on Machine and Deep Learning: A Review. Indones. J. Comput. Sci., 13.","DOI":"10.33022\/ijcs.v13i3.4059"},{"key":"ref_7","unstructured":"Cate, M. (2025, December 09). Cost-Sensitive Learning in Financial Fraud Detection Models. Available online: https:\/\/www.researchgate.net\/publication\/393569888_Cost-Sensitive_Learning_in_Financial_Fraud_Detection_Models."},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Hu, X., Chen, H., Chen, H., Liu, S., Li, X., Zhang, S., Wang, Y., and Xue, X. (2023). Cost Sensitive GNN-based Imbalanced Learning for Mobile Social Network Fraud Detection. arXiv.","DOI":"10.1109\/TCSS.2023.3302651"},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Qayoom, A., Khuhro, M.A., Kumar, K., Waqas, M., Saeed, U., Rehman, S.U., Wu, Y., and Wang, S. (2024). A novel approach for credit card fraud transaction detection using deep reinforcement learning scheme. PeerJ Comput. Sci., 10.","DOI":"10.7717\/peerj-cs.1998"},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"1325","DOI":"10.47392\/IRJAEM.2025.0216","article-title":"A Survey on Credit Card Fraud Detection using Deep Learning Model","volume":"3","author":"Jayalakshmi","year":"2025","journal-title":"Int. Res. J. Adv. Eng. Manag. (IRJAEM)"},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Li, Y., Wang, S., Ding, H., and Chen, H. (2023). Large Language Models in Finance: A Survey. arXiv.","DOI":"10.1145\/3604237.3626869"},{"key":"ref_12","unstructured":"Elkan, C. (2001, January 4\u201310). The foundations of cost-sensitive learning. Proceedings of the 17th International Joint Conference on Artificial Intelligence, IJCAI\u201901, Seattle, WA, USA."},{"key":"ref_13","doi-asserted-by":"crossref","unstructured":"Lin, T.Y., Goyal, P., Girshick, R., He, K., and Doll\u00e1r, P. (2017, January 22\u201329). Focal loss for dense object detection. Proceedings of the 2017 IEEE International Conference on Computer Vision (ICCV), Venice, Italy.","DOI":"10.1109\/ICCV.2017.324"},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Benchaji, I., Douzi, S., Ouahidi, B.E., and Jaafari, J. (2021). Enhanced credit card fraud detection based on attention mechanism and LSTM deep model. J. Big Data, 8.","DOI":"10.1186\/s40537-021-00541-8"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"16400","DOI":"10.1109\/ACCESS.2022.3148298","article-title":"A Neural Network Ensemble with Feature Engineering for Improved Credit Card Fraud Detection","volume":"10","author":"Esenogho","year":"2022","journal-title":"IEEE Access"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"14277","DOI":"10.1109\/ACCESS.2018.2806420","article-title":"Credit card fraud detection using AdaBoost and majority voting","volume":"6","author":"Randhawa","year":"2018","journal-title":"IEEE Access"},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Tanouz, D., Subramanian, R.R., Eswar, D., Parameswara Reddy, G.V., Ranjith Kumar, A., and Praneeth, C.H.V.N.M. (2021, January 6\u20138). Credit card fraud detection using machine learning. Proceedings of the 2021 5th International Conference on Intelligent Computing and Control Systems (ICICCS), Madurai, India.","DOI":"10.1109\/ICICCS51141.2021.9432308"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Alfaiz, N.S., and Fati, S.M. (2022). Enhanced credit card fraud detection model using machine learning. Electronics, 11.","DOI":"10.3390\/electronics11040662"},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Khalid, A.R., Gul, H., Nasir, J., Al-Issa, Y.I., and Al-Maroof, R.S. (2024). Enhancing credit card fraud detection: An ensemble machine learning approach. Big Data Cogn. Comput., 8.","DOI":"10.3390\/bdcc8010006"},{"key":"ref_20","unstructured":"Aminian, G., Elliott, A., Li, T., Wong, T.C.H., Dehon, V.C., Szpruch, L., Maple, C., Read, C., Brown, M., and Reinert, G. (2025). FraudTransformer: Time-Aware GPT for Transaction Fraud Detection. arXiv."},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Dang, T.K., Tran, T.C., Tuan, L.M., and Tiep, M.V. (2021). Machine Learning Based on Resampling Approaches and Deep Reinforcement Learning for Credit Card Fraud Detection Systems. Appl. Sci., 11.","DOI":"10.3390\/app112110004"},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Zhinin-Vera, L., Riofrio-Paz, F., Camacho-G., E., Brito-P., L., Chang, O., Valencia-Ramos, R., Velastegui, R., Pilliza, G.E., and Quinga-Socasi, F. (2020). Q-Credit Card Fraud Detector for Imbalanced Classification using Reinforcement Learning. Proceedings of the 12th International Conference on Agents and Artificial Intelligence\u2014Volume 1: ICAART, Valletta, Malta, 22\u201324 February 2020, SciTePress.","DOI":"10.5220\/0009156102790286"},{"key":"ref_23","unstructured":"Das, S., Saha, S., Coello, C.A., and Bansal, J.C. (2024). Deep Reinforcement Learning for Credit Card Fraud Detection. Proceedings of the Advances in Data-Driven Computing and Intelligent Systems, Springer."},{"key":"ref_24","unstructured":"Araci, D. (2019). FinBERT: Financial Sentiment Analysis with Pre-trained Language Models. arXiv."},{"key":"ref_25","doi-asserted-by":"crossref","unstructured":"Yang, X., Xu, J., Liu, S., Xu, B., and Li, Y. (2023). FinChain-BERT: A high-accuracy automatic fraud detection model based on NLP methods for financial scenarios. Information, 14.","DOI":"10.3390\/info14090499"},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Bhattacharya, I., and Mickovic, A. (2024). Accounting fraud detection using contextual language learning. Int. J. Account. Inf. Syst., 53.","DOI":"10.1016\/j.accinf.2024.100682"},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"24853","DOI":"10.1007\/s00521-024-10495-6","article-title":"Large language models in finance (finllms)","volume":"37","author":"Lee","year":"2025","journal-title":"Neural Comput. Appl."},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Kirkos, E., Boskou, G., Chatzipetrou, E., Tiakas, E., and Spathis, C. (2025, April 01). Exploring the Boundaries of Financial Statement Fraud Detection with Large Language Models. Available online: https:\/\/ssrn.com\/abstract=4842962.","DOI":"10.2139\/ssrn.4897041"},{"key":"ref_29","first-page":"27730","article-title":"Training language models to follow instructions with human feedback","volume":"35","author":"Ouyang","year":"2022","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_30","unstructured":"Ahn, M., Brohan, A., Brown, N., Chebotar, Y., Cortes, O., David, B., Finn, C., Fu, C., Gopalakrishnan, K., and Hausman, K. (2022). Do as I can, not as I say: Grounding language in robotic affordances. arXiv."},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Yang, Z., Yang, D., Dyer, C., He, X., Smola, A., and Hovy, E. (2016, January 12\u201317). Hierarchical attention networks for document classification. Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, San Diego, CA, USA.","DOI":"10.18653\/v1\/N16-1174"},{"key":"ref_32","unstructured":"Bahdanau, D., Cho, K., and Bengio, Y. (2016). Neural Machine Translation by Jointly Learning to Align and Translate. arXiv."},{"key":"ref_33","unstructured":"Lin, Z., Feng, M., dos Santos, C.N., Yu, M., Xiang, B., Zhou, B., and Bengio, Y. (2017). A structured self-attentive sentence embedding. arXiv."},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"291","DOI":"10.1016\/j.ejor.2021.05.028","article-title":"Instance-dependent cost-sensitive learning for detecting transfer fraud","volume":"297","author":"Baesens","year":"2022","journal-title":"Eur. J. Oper. Res."},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Peykani, P., Peymany Foroushany, M., Tanasescu, C., Sargolzaei, M., and Kamyabfar, H. (2025). Evaluation of Cost-Sensitive Learning Models in Forecasting Business Failure of Capital Market Firms. Mathematics, 13.","DOI":"10.3390\/math13030368"},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"529","DOI":"10.1038\/nature14236","article-title":"Human-level control through deep reinforcement learning","volume":"518","author":"Mnih","year":"2015","journal-title":"Nature"},{"key":"ref_37","unstructured":"Schulman, J., Wolski, F., Dhariwal, P., Radford, A., and Klimov, O. (2017). Proximal Policy Optimization Algorithms. arXiv."},{"key":"ref_38","unstructured":"Mnih, V., Badia, A.P., Mirza, M., Graves, A., Lillicrap, T., Harley, T., Silver, D., and Kavukcuoglu, K. (2016, January 19\u201324). Asynchronous methods for deep reinforcement learning. Proceedings of the International Conference on Machine Learning, New York, NY, USA."},{"key":"ref_39","doi-asserted-by":"crossref","unstructured":"Li, L., Chu, W., Langford, J., and Schapire, R.E. (2010, January 26\u201330). A contextual-bandit approach to personalized news article recommendation. Proceedings of the WWW \u201910: Proceedings of the 19th International Conference on World Wide Web, Raleigh, NC, USA.","DOI":"10.1145\/1772690.1772758"},{"key":"ref_40","first-page":"487","article-title":"FinFD-GCN: Using Graph Convolutional Networks for Fraud Detection in Financial Data","volume":"12","year":"2024","journal-title":"J. AI Data Min."},{"key":"ref_41","unstructured":"Brockman, G., Cheung, V., Pettersson, L., Schneider, J., Schulman, J., Tang, J., and Zaremba, W. (2016). OpenAI Gym. arXiv."},{"key":"ref_42","first-page":"1","article-title":"Stable-Baselines3: Reliable reinforcement learning implementations","volume":"22","author":"Raffin","year":"2021","journal-title":"J. Mach. Learn. Res."},{"key":"ref_43","doi-asserted-by":"crossref","first-page":"139","DOI":"10.1016\/j.knosys.2017.05.001","article-title":"Mining corporate annual reports for intelligent detection of financial statement fraud: A comparative study of machine learning methods","volume":"128","author":"Hajek","year":"2017","journal-title":"Knowl. Based Syst."},{"key":"ref_44","doi-asserted-by":"crossref","unstructured":"Craja, P., Kim, A., and Lessmann, S. (2020). Deep learning for detecting financial statement fraud. Decis. Support Syst., 139.","DOI":"10.1016\/j.dss.2020.113421"},{"key":"ref_45","doi-asserted-by":"crossref","first-page":"178","DOI":"10.30574\/wjarr.2025.26.2.1586","article-title":"Application of LLMs to Fraud Detection","volume":"26","author":"Malingu","year":"2025","journal-title":"World J. Adv. Res. Rev."},{"key":"ref_46","doi-asserted-by":"crossref","first-page":"175829","DOI":"10.1109\/ACCESS.2024.3502542","article-title":"A Hybrid Deep Learning Ensemble Model for Credit Card Fraud Detection","volume":"12","author":"Ileberi","year":"2024","journal-title":"IEEE Access"},{"key":"ref_47","doi-asserted-by":"crossref","first-page":"1985","DOI":"10.1007\/s10796-022-10346-6","article-title":"Fraud detection in mobile payment systems using XGBoost-based frameworks","volume":"25","author":"Hajek","year":"2023","journal-title":"Inf. Syst. Front."},{"key":"ref_48","doi-asserted-by":"crossref","first-page":"211","DOI":"10.38193\/IJRCMS.2025.7217","article-title":"Graph-Enhanced Transformer Network for Fraud Detection in Digital Banking: Integrating GNN and Self-Attention for End-to-End Transaction Analysis","volume":"7","author":"Bolla","year":"2025","journal-title":"Int. J. Res. Commer. Manag. Stud."},{"key":"ref_49","doi-asserted-by":"crossref","unstructured":"Agomuo, O.C., Uzoma, A.K., Khan, Z., Otuomasirichi, A.I., and Muzamal, J.H. (2025, January 3\u20135). Transparent AI for Adaptive Fraud Detection. Proceedings of the 2025 19th International Conference on Ubiquitous Information Management and Communication (IMCOM), Bangkok, Thailand.","DOI":"10.1109\/IMCOM64595.2025.10857433"},{"key":"ref_50","doi-asserted-by":"crossref","first-page":"426","DOI":"10.1109\/OJCS.2025.3543450","article-title":"FraudGNN-RL: A Graph Neural Network with Reinforcement Learning for Adaptive Financial Fraud Detection","volume":"6","author":"Cui","year":"2025","journal-title":"IEEE Open J. Comput. Soc."}],"container-title":["Algorithms"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1999-4893\/18\/12\/792\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,16]],"date-time":"2025-12-16T08:59:02Z","timestamp":1765875542000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1999-4893\/18\/12\/792"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,15]]},"references-count":50,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2025,12]]}},"alternative-id":["a18120792"],"URL":"https:\/\/doi.org\/10.3390\/a18120792","relation":{},"ISSN":["1999-4893"],"issn-type":[{"value":"1999-4893","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,15]]}}}