Wednesday webinar at one: Understanding risks from deep learning frameworks

Wednesday 13 April 2022 | 13:00 - 13:40 | Webinar

This session is brought to you by CFA UK’s Data Science working group

Deep learning has proved to be a powerful tool among data scientists. Deep Learning models power autonomous cars to robot traders and are rapidly being adopted for credit assessment, lending and more.

This webinar will introduce to the audience research about the risks emerging from deep learning frameworks and how these could be monitored using an automated technique.

This session will discuss:

  • What are deep learning models and how are they different from traditional analytical methods?
  • Channels through which deep learning models can cause systemic and idiosyncratic risks
  • Idiosyncratic risks – Explainability, robustness
  • Systemic -  Herding and monocultures
  • Monitoring
 Where recordings are made, these are a member benefit that are accessed through the member-only platform, CFA UK Discover.


Registration: 12:55

Event: 13:00 - 13:40

CPD Points: 0.75


Rishabh Kumar, Research and Development Data Scientist, Prudential Regulatory Authority/Bank of England

Rishabh Kumar is a Research and Development Data Scientist at the Prudential Regulatory Authority/Bank of England. His research work involves understanding and exploiting benefits of AI and ML to drive forward the Bank’s RegTech Agenda supporting the PRA’s remit to regulate c1500 banks, building societies and insurance companies.

His most recent research work investigates whether certain machine learning models, if incorrectly deployed, could lead to unmitigated risks for the firm and wider contagion impacts to the wider sector.

Prior to working at the bank, Rishabh worked for a behavioural data science consultancy firm advising a range of companies on how they can use psychology and data science to maximise insights.

Rishabh has worked at research roles at Warwick University and Sheffield University. He is a member of Alan Turing Institute’s Behavioural data science group.

Click here to view our event terms and conditions.