Conference Day One
- Machine learning and deep learning applications in quantitative finance and risk management
- Practitioners’ case studies
- Research and development in deep learning
Executive Director, Morgan Stanley
Marcelo Labre: Executive Director, Morgan Stanley
Professor and Dept. Chair of FRE Tandon, New York University
Peter Carr: Professor and Dept. Chair of FRE Tandon, New York University
Dr. Peter Carr is the Chair of the Finance and Risk Engineering Department at NYU Tandon School of Engineering. He has headed various quant groups in the financial industry for the last twenty years. He also presently serves as a trustee for the National Museum of Mathematics and WorldQuant University. Prior to joining the financial industry, Dr. Carr was a finance professor for 8 years at Cornell University, after obtaining his Ph.D. from UCLA in 1989. He has over 85 publications in academic and industry-oriented journals and serves as an associate editor for 8 journals related to mathematical finance. He was selected as Quant of the Year by Risk Magazine in 2003 and Financial Engineer of the Year by IAQF/Sungard in 2010. From 2011 to 2014, Dr. Carr was included in Institutional Investor’s Tech 50, an annual listing of the 50 most influential people in financial technology.
In the 2 years Dr. Carr been FRE dept. chair, applications increased from 1,300 per year to 1,900 per year. The number of FRE Masters students in residence was the highest in any 2-year period. For the incoming 2018 class, current verbal GRE is 169/170 and GPA is 3.82. FRE moved up in Quantnet rankings both years. An online summer course was initiated last summer and an on-campus bootcamp will be initiated this summer. Six electives on machine learning in finance were introduced. The distance learning room will become operational this summer.
- Modern Data Analysis
- Times Series Models Univariate
- Linear Factor Models
- Multivariate Time Series
- Modern Financial Engineering
- Long Short Term Memory Networks
Miquel Noguer Alonso:
Adjunct Assistant Professor, COLUMBIA UNIVERSITY
Miquel Noguer Alonso: Adjunct Assistant Professor, COLUMBIA UNIVERSITY
Miquel Noguer i Alonso is a financial markets practitioner with more than 20 years of experience in asset management, he is currently working for UBS AG (Switzerland). He worked as a CFO and CIO for a European bank from 2000 to 2006. He started his career at KPMG.
He is Adjunct Assistant Professor at Columbia University teaching Asset Allocation, Big Data in Finance, Fintech and Hedge Fund Professor at ESADE. He received an MBA and a Degree in business administration and economics in ESADE in 1993. In 2010 he earned a PhD in quantitative finance with a Summa Cum Laude distinction (UNED – Madrid Spain). He also holds the Certified European Financial Analyst diploma ( 2000 ).
His research interests range from asset allocation, big data to algorithmic trading and fintech. His academic collaborations include a visiting scholarship in Columbia University in 2013 in the Finance and Economics Department, in Fribourg University in 2010 in the mathematics department, and presentations in Indiana University, ESADE, London Business School, CAIA Association, AFI and several industry seminars.
In discrete time, option hedging and pricing amount to sequential risk minimization. In particular, a discrete-time version of the Black-Scholes-Merton (BSM) option pricing model can be formulated as a problem of dynamic Markowitz optimization of an option replicating (hedge) portfolio made of an underlying stock and cash. This talk shows how this problem can be approached using Reinforcement Learning (RL). Once the problem is posed as an RL problem, option pricing and hedging can be done without any model for the underlying stock dynamics, using instead model-free, data-driven RL methods such as Q-learning and Fitted Q Iteration. As a result, both option price and hedge are obtained by a well-defined and converging maximization problem that uses only market prices and option trading data (inter-temporal re-hedges and hedge losses in the replicating portfolio) to find the optimal option hedge and price. The model can learn when re-hedges in data are suboptimal/noisy, or even purely random. This means, in particular, that our RL model can learn the BSM model itself, if the world is according to BSM.
Computationally, the RL-based option pricing model is very simple, as it uses only basic linear algebra and linear regressions to compute the option price and hedge. The only tunable parameters in this approach are parameters defining the optimal hedge and price themselves. This approach does not need any model calibration (as there is no model anymore), and it automatically solves the volatility smile problem of the BSM model. We also discuss some extensions of this approach, including in particular an Inverse Reinforcement Learning setting, where inter-temporal losses from re-hedges are unobservable.
Research Professor of Financial Machine Learning, NYU Tandon School of Engineering
Igor Halperin: Research Professor of Financial Machine Learning, NYU Tandon School of Engineering
Igor Halperin is currently an Adjunct Professor of Financial Machine Learning at the NYU Tandon School of Engineering. Prior to that, he was an Executive Director of Quantitative Research at JPMorgan Chase where he focused on the research and development of predictive and statistical models and machine learning methods for modeling risk of financial portfolios.
He has authored a number of publications on quantitative finance, and is a frequent speaker at financial conferences. Dr. Halperin has a Ph.D. in theoretical physics from Tel Aviv University, and M.Sc. in nuclear physics from St. Petersburg State Technical.
Managing Director, Citi Institutional Clients Group
Terry Benzschawel: Managing Director, Citi Institutional Clients Group
Terry Benzschawel is a Managing Director in Citigroup’s Institutional Clients Business. Terry heads the Credit Trading Analysis group which develops and implements quantitative tools and strategies for credit market trading and risk management, both for Citi’s clients and for in-house applications. Some sample tools include models of corporate default and recovery values, relative value of corporate bonds, loans, and credit default swaps, credit portfolio optimization, credit derivative trades, capital structure arbitrage, measuring and hedging liquidity risk, and cross-credit-sector asset allocation.
After six years of post-doctoral research in academia and industry and two years in consumer banking, Terry began his investment banking career in at Salomon Brothers in 1992. Terry built models for proprietary arbitrage trading in bonds, currencies and derivative securities in Salomon’s Fixed Income Arbitrage Group. In 1998, he moved to the Fixed Income Strategy department as a credit strategist with a focus on client-oriented solutions across all credit markets and has worked in related roles since then. Terry was promoted to Managing Director at Citi in 2008.
Terry received his Ph.D. in Experimental Psychology from Indiana University (1980) and his B.A. (with Distinction) from the University of Wisconsin (1975). Terry has done post-doctoral fellowships in Optometry at the University of California at Berkeley and in Ophthalmology at the Johns Hopkins University School of Medicine and was a visiting scientist at the IBM Thomas J. Watson Research Center prior to embarking on a career in finance. He currently serves on the steering committees of the Masters of Financial Engineering Programs at the University of California at Berkeley and the University of California at Los Angeles and Carnegie Mellon University’s Computational Finance Program.
Terry is a frequent speaker at industry conferences and events and has lectured on credit modelling at major universities. In addition, he has published over a dozen articles in refereed journals and is author of CREDIT MODELING: FACTS, THEORIES AND APPLICATIONS. In addition, Terry has been the instructor for courses in credit modelling for Incisive Media and the Centre for Finance Professionals. Finally, Terry has taught a course on credit modelling at Russia’s Sberbank in Moscow.
Many publicly available examples of time series prediction with neural networks use fake or
random data. Other examples, particularly in finance, present poorly performing models. It is very hard to learn good practices when only presented with toy examples. Instead, this talk aims to teach the full process of using a neural network for time series prediction by walking through a real problem from start to finish.
We will begin by explaining the concrete problem we would like to solve and how to frame our problem in a way that we can model. Once we understand our problem, we will discuss how to collect the needed data. We will discuss the process of reducing our input data into important features for the model to consume. We will then learn how to use Keras to implement our neural network. Once we have a working model, we will cover some tricks to improve its performance.
At every step, we will cover problems faced while working on this model. We will show how to use data visualization to aid in model development and catch problems early. We will also cover tips for using numpy to work with time series data efficiently.
By the end of the talk, audience members will:
- Know how to frame a problem in a way that a neural network can model
- Know how to think about feature selection
- Be familiar with the Keras API for time series predictions
- Understand that the hardest problems come before you even get to Keras
Applying Machine Learning to Evaluate Systemic Risk and Contribution of Individual SIFIs
Senior Advisor, Deloitte
Ksenia Shnyra: Senior Advisor, Deloitte
- Luis Cota: Data Scientist, Thalesians
- Miquel Noguer Alonso: Adjunct Assistant Professor, Columbia University
- Igor Halperin: Research Professor of Financial Machine Learning, NYU Tandon School of Engineering
- Marcelo Labre: Executive Director, Morgan Stanley
- Peter Olausson: CEO, COGNITUUM Artificial General Intelligence
- What is the current state of utilisation of machine learning in finance?
- What are the distinct features of machine learning problems in finance compared to other industries?
- What are the best practices to overcome these difficulties?
- What’s the evolution of a team using machine learning in terms of day to day operations?
- What is a typical front office ‘Quant’ skillset going to look like in three to five years time?
- How do we deal with model risk in machine learning case?
- How is machine learning expected to be regulated?
- What applications can you list among its successes?
- How much value is it adding over and above the “classical” techniques such as linear regression, convex optimisation, etc.?
- Do you see high-performance computing (HPC) as a major enabler of machine learning?
- What advances in HPC have caused the most progress?
- What do you see as the most important machine learning techniques for the future?
- What are the main pitfalls of using Machine Learning currently in trading strategies?
- What new insights can Machine Learning offer into the analysis of financial time series?
- Discuss the potential of Deep Learning in algorithmic trading?
- Do you think machine learning and HPC will transform finance 5-10 years from now?
- If so, how do you envisage this transformation?
- Can you anticipate any pitfalls that we should watch out for