After successful completion of the course, students are able to...

• describe several types of neural networks (recurrent, convolutional), and to discuss their topology
• to apply stochastic gradient descent including curvature
• to implement the resulting algorithms via Keras and Tensorflow
• to apply the resulting Deep Learning technology on problems of financial and actuarial mathematics

• required foundations of linear algebra and Bayesian statistics
• recurrent neural networks and their application in importance sampling for valuation of derivatives, limit order book, solvency capital requirement, life insurance risk estimation
• convolutional neural networks and their application for financial time series

• blackboard presentation
• conceptional mathematical proofs
• programming in Python based on the Keras/Tensorflow platform

oral exam