Committee Meeting, Free for Members, Group Meeting, Interest Group, Interest Group Meetings, Quantitative Investing, Virtual Events & Programming

Multistep forecasting for limit order books inspired by machine translation algorithms

October 8, 2021 | 12:00 PM - 1:30 PM

Share Event

Loading Events

Wait! A Note on Registration:

We’ve launched Cvent—our new events platform!

Registration for any event with a start date after Sept. 28 now requires a CFA Institute account.

I don’t have a CFA Institute account

  • No problem! You’ll have the chance to create one prior to registration.

I already have a CFA Institute account

  • Great! Be sure to use your existing credentials at registration.
Cvent Transition Guide


Initially we review the use of deep learning methods from image analysis and natural language processing applied to the problem of forecasting financial time-series. We extend these methods with state-of-the-art techniques from machine translation and thereby propose a novel multi-horizon forecasting model for limit order book data. Our empirical findings demonstrate that this new models outperforms existing methods significantly, however it is computationally expensive. We overcome this computational bottleneck by employing a cutting-edge hardware accelerator called an Intelligence Processing Unit (IPU). Our results demonstrate that IPUs can offer order of magnitude speed-ups compared with even Graphical Processing Units (GPUs) – the enabling technology that underpins the Deep Learning revolution of the last decade.

Quantitative Investing Group

Guest Speaker

Stefan Zohren
Professor of Machine Learning in Finance
Oxford-Man Institute and Scientific Advisor at MAN Group

Stefan Zohren is a Faculty Member at the Oxford-Man Institute of Quantitative Finance, an Associate at the Oxford Internet Institute and a Mentor in the FinTech stream at the Creative Destruction Lab at Said Business School, all at the University of Oxford. He is also a Fellow at the Alan Turing Institute and a member of the European Laboratory for Learning and Intelligent Systems. Outside of academia, Stefan works as a Scientific Advisor at Man Group. He has extensive consulting experience in leading machine learning projects in application domains such as finance and healthcare with clients ranging from listed companies to SMEs and startups. Stefan is also a frequent speaker on AI in finance at academic conferences, as well as industry panels and corporate events.