• Identify risk areas that can be modeled with data and explain what is required to do so
• Assess the limitations of data driven risk management and identify ways how to overcome them if possible
• Demonstrate different modeling approaches for risk and describe their attached limitations
• Calculate such risk models in R or Python for simple data examples
• Discuss the governance element and explain ethical considerations related to such quantitative modeling’s e.g. in a finance application

The global business environment is influenced by many emergent risk factors that can lead to unforeseen events and changes in business dynamics. Besides dramatic changes such as financial crisis, political conflicts, competitive disruption, natural disasters, pandemics, etc. there are also uncertainties that can be modeled using modern algorithm based approaches. This introduces new potentially dramatic threats or challenges but at the same time offers opportunities to be explored and exploited.

The use of statistical models in computer algorithms allows computers to make decisions and predictions, and to perform tasks that traditionally require human cognitive abilities. Machine learning is the interdisciplinary field at the intersection of statistics and computer science, which develops such statistical models and computer algorithms. It underpins many modern technologies, such as speech recognition, Internet search, bioinformatics and computer vision for example Amazon’s recommender system, Google’s driverless car and the most recent imaging systems for cancer diagnosis make use of Machine Learning technology.

This course on quantitative risk management with certain elements of Machine Learning will explain how to build systems that learn and adapt using real-world applications to detect and manage risks in a corporate setting. Some of the topics to be covered include linear regression, logistic regression, deep neural networks, clustering, and so forth. The course will be project-oriented with emphasis placed on writing software scripts of learning algorithms applied to real-world problems, in particular, Credit Risk, Collections Management and Fraud Detection.

Besides the technical elements, the course will touch upon the requirements that are set for an organization to use ML in such highly relevant application. It will have discussion sessions that touch upon the limitations and necessities around these sort of models.

The course attempts to advance critical thinking on quantitative risk management in open class discussions and group exercises analyzing (often well-known) situations that demonstrate shortcomings and challenges off certain methods. It outlines modern quantitative risk management and considers how leaders can deal (more) effectively with contemporary environments. The shortcomings of current practices are addressed with an intent to develop effective approaches to deal with risks we face today.

[1] Aziz, S., & Dowling, M. (2019). Machine learning and AI for risk management. In Disrupting finance (pp. 33-50). Palgrave Pivot, Cham.

[2] Leo, M., Sharma, S., & Maddulety, K. (2019). Machine learning in banking risk management: A literature review. Risks, 7(1), 29.

[3] Van Liebergen, B. (2017). Machine learning: a revolution in risk management and compliance?. Journal of Financial Transformation, 45, 60-67.

[4] Chandrinos, S. K., Sakkas, G., & Lagaros, N. D. (2018). AIRMS: A risk management tool using machine learning. Expert Systems with Applications, 105, 34-48.

[5] Fécamp, S., Mikael, J., & Warin, X. (2019). Risk management with machine-learning-based algorithms. arXiv preprint arXiv:1902.05287.

[6] Dionne, G. (2013). Risk management: History, definition, and critique. Risk management and insurance review, 16(2), 147-166.