ENCORE - Engaging New Controversial Original Rigorous Entertaining
Human input is critical to model design & management
Seminar by David Jessop
14th July 2020 – On-Line
In this seminar David Jessop will discuss both the need for and value of human input to the development and management of complex computer based models, especially those used in finance, and the need to understand and interpret their output.
There is a philosophical view which says that machine learning is “the end of theory”. No longer do we have to build theoretical constructs and make assumptions (many of which are or may be wrong) to support our modelling – we just find enough data, throw it into any “suitable” algorithm and as if by magic the answer will appear.
This seminar will present a different view. Divorcing modelling from models of the world can work in certain circumstances, but these circumstances rarely if ever really exist in financial markets. So, humans, alongside machines, are still needed to develop investment strategies which can be interpreted and explained to asset owners – both when they have produced good results and more importantly when they have not performed as had been hoped!
Head of Investment Risk, EMEA, at Columbia Threadneedle Investments
David Jessop has responsibility for overseeing the independent investment risk management process for all portfolios managed in the EMEA region.
Before joining Columbia Threadneedle Investments, David was the Global Head of Quantitative Research at UBS. Over his 17 years at UBS his research covered many topics but in particular he concentrated on risk analysis, portfolio construction and more recently cross asset factor investing / the application of machine learning and Bayesian techniques in investment management. Prior to this he was Head of Quantitative Marketing at Citigroup. David started his career at Morgan Grenfell, initially as a derivative analyst and then as a quantitative portfolio manager.
David has a MA in Mathematics from Trinity College, Cambridge.