Waterloo tests AI for use in financial sector

The Univ of Waterloo has developed software with AI algorithms that can help financial institutions justify the decisions they make as they comply with regulations.  Waterloo’s deep-learning algorithms are able to process and analyze rich data, and detect patterns that even users can’t tell how the analyses were made, researchers say. To test the algorithms, the research team used data from the S&P 500 stock index and taught the AI system to predict next-day outcomes. The researchers then studied different factors – such as index high, low, open and close levels, and trading volume – to determine how predictions were arrived at using AI. Researchers say this experiment can help in real-life applications, such as when the financial institutions are asked to justify their decisions in approving mortgages or setting insurance premiums. Waterloo researchers say they expect to deploy the software for field trials within the year. The research team is led by Devinder Kumar, a PhD candidate in systems design engineering at Waterloo. Kumar collaborated with professors Alexander Wong of Waterloo and Graham Taylor of the Univ of Guelph.