The Centre for Robust Inference in a Digital Economy (RIDE) was approved by Royal Holloway in October 2016. Its main area of research is robust inference as the basis for robust decision making in finance, economics, and other social sciences.
The emphasis is on the opportunities and challenges resulting from the digitalisation of society, e.g. big data availability, flash crashes and abuse in financial markets caused by algorithms,data integrity/fabrication of electronic surveys, etc.
We are interested in methods to conduct inference and making decisions which are nearly optimal in the face of considerable uncertainty and in ways to make various social networks of decision makers (such as financial markets) less fragile and prone to users' abuse.
Developing robust methods of prediction is also covered,as prediction can be fruitfully considered as a special case of both inference and decision making.
The basis of inference is data, and therefore we are also interested in data collection data processing, including data analysis and big-data technologies (such as MapReduce)and data distribution.
Here are some of our goals:
- Developing robust methods of inference
- Developing robust methods of decision making
- Developing methods of making various social networks (including financial markets) less fragile
- Defining and testing various regulations and policy mechanisms to make society immune from treats that arise implicitly or explicitly in the digital economy using data and economic experiments.
We have a RIDE sponsored event 'Advances in Factor Models for Finance' at Senate House in London on Friday 7 December at 2pm
Our staff and research students
RIDE is a joint venture between our Departments of Computer Science and Economics.
Professor Alessio Sancetta (Economics)
Professor Vladimir Vovk (Computer Science)
Dr Gregory Chockler (Computer Science)
Professor Francesco Feri (Economics)
Professor Alex Gammerman (Computer Science)
Dr Yuri Kalnishkan (Computer Science)
Dr Zhiyuan Luo (Computer Science)
Dr Michael Naef (Economics)
Professor Michael Spagat (Economics)
Professor Kostas Stathis (Computer Science)
Professor Chris Watkins (Computer Science)
Professor Klaus Ritzberger (Economics)
Dr Jungyoon Lee (Economics)
Miss Raisa Dzhamtyrova, supervised by Yuri Kalnishkan
Mr Valery Manokhin, supervised by Vladimir Vovk and Alessio Sancetta
Mr Luca Mucciante, supervised by Alessio Sancetta and Vladimir Vovk
Mr Philip Nadler, supervised by Mike Spagat and Zhiyuan Luo
Mr Ivan Petej, supervised by Vladimir Vovk
The following are key publications on the topics of robust inference and decision making by RIDE's members. More publications are to appear shortly.
Publications by Alex Gammerman
Gammerman, A. and V. Vovk (2007). Hedging predictions in machine learning (with discussion). Computer Journal, Vol. 50, No. 2, p. 151-163.
Nouretdinov, I., T. Bellotti, and A. Gammerman (2014). Diagnostic and prognostic by conformal predictors. In: Conformal Predictions for Reliable Machine Learning: Theory, Adaptations and Applications (ed. by V. Balasubramanian, S.-S. Ho, and V. Vovk), p. 217–230. Elsevier, Amsterdam.
Publications by Alessio Sancetta
Sancetta, A. (2012) Universality of Bayesian predictions. 'Bayesian Analysis, Vol. 7, No 1, p. 1-36, p. 45-46 (rejoinder)'. This article establishes finite sample bounds for the loss incurred using Bayesian predictions in a variety of settings, including almost arbitrary loss functions, model averaging, predictions in a non-stationary environment and under model misspecification.
Kurov, A., A. Sancetta, G.Strasser, and M. Halova Wolfe (2016) Price drift before U.S. macroeconomic news: private information about public announcements? This article shows that trading on information leakage of US economic announcements could have taken place since 2008 resulting in profits possibly in the order of hundred of millions in US dollars.
Publications by Vladimir Vovk
Shafer, G. and V. Vovk (2001) Probability and Finance: It's Only a Game! New York : Wiley-Interscience. This book introduces a robust approach to mathematical finance and the foundations of probability. A new edition is forthcoming.
Vovk, V. (2012) Continuous-time trading and the emergence of probability. 'Finance and Stochastics, Vol. 16, No. 4, p. 561 - 609'. This article introduces a probability-free approach to the study of security prices in idealised financial markets.
The main contribution from RIDE to this project is the extension of the theory of algorithmic complexity to probability-free prediction.
Market manipulation and abuse in cryptocurrencies and other electronic markets
The main contribution from RIDE to this project is the definition of market manipulation in a way that can be tested using empirical data. The goal is infer the extent to which electronic markets are prone to illegal practices such as spoofing.
Online Machine Learning for Effective Market Making Pricing and Risk Management Hedging Strategies
This is a joint project with Algolabs: see http://algolabs.com/PhD.htm.