UCL Institute of Finance & Technology



An intellectual hub of innovative ideas and critical analysis of AI modelling in risk management

We live in an era where scientific fields are overwhelmingly impacted by AI, becoming ever more reliant on it, and finance is no exception. Although the use of AI in financial modelling has already begun and the prospects are promising, the complexity of AI algorithms entails that it is not clear how and to what extent such algorithms will reshape the world of finance. To face these challenges, now more than ever, a thorough analysis and investigations of these models with innovative ideas and solutions are required.


The Artificial Intelligence Risk Laboratory (AIRiskLab) is committed to leveraging AI modelling and algorithms in risk management and study their long- and medium-term impacts on financial stability and sustainability, based on a broad interpretation of risk in finance. AIRiskLab will be at the forefront of this challenge and will provide a blend of cutting edge solutions and advice, as an intellectual hub of innovative ideas and critical analysis of AI modelling in risk management.

The lab was founded and is currently led by Dr Ramin Okhrati, a full-time faculty member of the Institute of Finance & Technology (IFT). 

Academic and Advisory Members

Dr Ramin Okhrati is the head of the AIRiskLab at IFT. He is also a full-time faculty member of the Institute and the Programme Director of the MSc  Banking and Digital Finance. He has extensive experience in theoretical and computational finance (driven by AI and machine learning methods) and financial technologies both from industrial (through consultancy and collaboration) and academic perspectives. His background is in mathematical finance; he holds a BSc, an MSc in Mathematics, and a PhD in Applied Probability with concentration in actuarial and finance. Beside his academic position, he also collaborates with the Bank of England as a senior researcher. He has publications in leading international journals and a wide range of research interests, in particular, in machine learning, digital finance, credit risk, insurance, stochastic analysis, and applied probability. Before joining UCL, he worked at the mathematical sciences of the University of Southampton, UK, and a held a one-year postdoctoral fellowship at the Financial and Actuarial Mathematics group of Vienna University of Technology in Vienna, Austria.

Prof Francesca Medda is Director and Professor of Applied Economics and Finance at UCL IFT. Professor Medda's key research areas are digital finance, impact & sustainable finance, and urban & infrastructure investments.

Dr Aldo Lipani is a Lecturer (Assistant Professor) in Machine Learning. Aldo is a member of the SpaceTimeLab led by Prof. Tao Cheng and of the Web Intelligence Group led by Prof. Emine Yilmaz. Previously, Aldo was a postdoctoral research at UCL. Aldo received his Ph.D. in Computer Science from the TU Wien (Austria) with the dissertation titled: “On Biases in Information Retrieval Models and Evaluation” under the supervision of Prof. Allan Hanbury and Dr. Mihai Lupu. He has furthered his studies at the: National Institute of Standard and Technologies (NIST) in Gaithersburg, Microsoft Research Cambridge, University of Glasgow, University of Amsterdam, and National Institute of Informatics (NII) in Tokyo.

PhD students

Raad Khraishi is a PhD student at IFT. He has several years of industry experience in data science and finance and is currently Lead Data Scientist within the Data Science & Innovation team at NatWest Group.

Pin Ni is a PhD student at IFT. His current research interests include Deep Learning, Natural Language Processing, Knowledge Graph, FinTech, Medical Informatics, etc. He has served as a reviewer in NLP, AI, Information Retrieval, Database, FinTech, Medical Informatics and other computer science fields’ top venues. He has also served as co-PI, co-I, convener and core member for several government and enterprise key research projects in different countries.

Zihao Liu is a PhD student and teaching assistant at IFT. He holds double Masters’ degrees in Financial Engineering and Banking and Digital Finance. With strong interests in ESG (Environment, Social and Governance) and green finance, Zihao’s current research interests include ESG investing, climate change Finance and Green Finance by the supporting of Machine Learning. ESG related topics are studied by many academic researchers but there is still no consistent conclusion on ESG investing. His aim is to enhance the field of ESG research and extend the research to the climate changing financial market with new solutions. Zihao is also a FRM charter holder with internship in Quantexa as a data analyst.

Viktor Kazakov is a treasury professional with more than 11 years of experience working in the private and IFI sectors. He has vast experience in varied areas such as asset liability management, structuring of interest rate and cross currency swaps, fixed income portfolio management and project finance. Viktor currently works at the European Bank for Reconstruction and Development in London where he has the dual role of being responsible for building in-house web and VBA applications as well as for advising internal and external clients and structuring deals. Also, he is a CFA Charter holder with a post-master degree from Sciences Po, Paris and a Master's Degree from the University of Sofia.
Daniil Bargman is a PhD student at IFT. He is also a director at the Chief Investment Office of UBS Global Wealth Management with nearly 10 years of professional experience in the investment industry. Daniil's specialities are macroeconomic scenario analysis, portfolio risk management, and multi-asset investment research. Daniil holds a dual MSc in Finance and International Management from the Stockholm School of Economics.


Offline Reinforcement Learning for Contractual Pricing
Sponsor: NatWest Group
Lead Investigators: Raad Khraishi, Dr Ramin Okhrati, Prof Francesca Medda
Start date: September 2021
Project summary: Raad is investigating offline reinforcement learning approaches to pricing contractual products. Traditional online reinforcement learning is impractical in many pricing applications given potential for mispricing as an agent learns from live interactions through trial and error. He is currently exploring offline techniques that allow an agent to learn a pricing policy using historical data with online fine-tuning.

A machine learning approach to support decisions in insider trading detection
Sponsor Commissione Nazionale per le Società e la Borsa (Consob)
Lead Investigators: Piero Mazzarisi, Adele Ravagnani, Paola Deriu, Fabrizio Lillo, Francesca Medda, Antonio Russo
Start date: May 2021
End date: September 2022
Project summary: Identifying market abuse activity from data on investors' trading activity is very challenging both for the data volume and for the low signal to noise ratio. Here we propose two complementary unsupervised machine learning methods to support market surveillance aimed at identifying potential insider trading activities. The first one uses clustering to identify, in the vicinity of a price sensitive event such as a takeover bid, discontinuities in the trading activity of an investor with respect to his/her own past trading history and on the present trading activity of his/her peers. The second unsupervised approach aims at identifying (small) groups of investors that act coherently around price sensitive events, pointing to potential insider rings, i.e. a group of synchronised traders displaying strong directional trading in rewarding position in a period before the price sensitive event. As a case study, we apply our methods to investor resolved data of Italian stocks around takeover bids.
Paper:  https://arxiv.org/pdf/2212.05912.pdf

Recent Publications

Wattanawongwan, S., Mues, C., Okhrati, R., Choudhry, T. and So, M.C., 2023. Modelling credit card exposure at default using vine copula quantile regression. European Journal of Operational Research.

Hamill, C.B., Khraishi, R., Gherghel, S., Lawrence, J., Mercuri, S., Okhrati, R. and Cowan, G.A., 2023. Agent-based Modelling of Credit Card Promotions. arXiv preprint arXiv:2311.01901.

Khraishi, R. and Okhrati, R., 2023. Simple Noisy Environment Augmentation for Reinforcement Learning. arXiv preprint arXiv:2305.02882.

Ni, P., Yuan, Q., Khraishi, R., Okhrati, R., Lipani, A., & Medda, F. (2022). Eigenvector-based Graph Neural Network Embeddings and Trust Rating Prediction in Bitcoin Networks. Proceedings of the 3rd ACM International Conference on AI in Finance. USA: ACM. doi:10.1145/3533271.3561793

Khraishi, R and Okhrati, R, 2022, Offline Deep Reinforcement Learning for Dynamic Pricing of Consumer CreditProceedings of the 3rd ACM International Conference on AI in Finance, ICAIF 2022, 325-333. doi:10.1145/3533271.3561682

Wattanawongwan, S., Mues, C., Okhrati, R., Choudhry, T., & So, M. C. (2022). A mixture model for credit card exposure at default using the GAMLSS frameworkInternational Journal of Forecasting. doi:10.1016/j.ijforecast.2021.12.014

Mercuri, S., Khraishi, R., Okhrati, R., Batra, D., Hamill, C., Ghasempour, T. and Nowlan, A., 2022. An Introduction to Machine Unlearning. arXiv preprint arXiv:2209.00939.

Ni, P., Okhrati, R., Guan, S., & Chang, V. (2022). Knowledge Graph and Deep Learning-based Text-to-GQL Model for Intelligent Medical Consultation Chatbot. Information Systems Frontiers. doi:10.1007/s10796-022-10295-0

    Okhrati, R., & Karpathopoulos, N. (2021). Local Risk Minimization of Contingent Claims Simultaneously Exposed to Endogenous and Exogenous Default TimesInternational Journal of Theoretical and Applied Finance. doi:10.1142/s0219024921500333