A1411
Title: A time series approach to explainability for neural nets with applications to risk-management and fraud detection
Authors: Marc Wildi - Zurich University (Switzerland) [presenting]
Branka Hadji Misheva - ZHAW Zurich University of Applied Sciences (Switzerland)
Abstract: Artificial intelligence (AI) is creating one of the biggest revolutions across technology-driven application fields. The finance sector offers many opportunities for significant market innovation, and yet broad adoption of AI systems heavily relies on trust in their outputs. Trust in technology is enabled by understanding the rationale behind the predictions made. To this end, the concept of eXplainable AI (XAI) emerged, introducing a suite of techniques attempting to explain to users how complex models arrived at a certain decision. For cross-sectional data, classical XAI approaches can lead to valuable insights into the model's inner workings. Still, these techniques generally cannot cope with longitudinal data (time series) in dependence structure and non-stationarity. A novel XAI technique is proposed for deep learning methods (DL) which preserves and exploits the natural time ordering of the data. Simple applications to financial data illustrate the potential of the new approach in the context of risk management and fraud-detection.
