Event Agenda

Agenda

Agenda: Machine Leaning in Finance - New York

Day one: Tuesday, March 17, 2020

08:3009:00

Registration and refreshments

08:30 - 09:00

09:0010:30

Introduction to machine learning and early financial applications

09:00 - 10:30

  • Big data versus machine learning 

  • Decision trees 

  • Introduction to neural networks 

    • The perceptron evolution of the artificial neuron 

    • Activation functions 

    • Backpropagation and gradient descent 

    • Adaptive learning mechanisms 

  • Optimization and regularization  

    • Regularization, optimization and deep learning

    • L1 and L2 regularization 

    • Optimization algorithms 

    • Adaptive learning methods 

  • Early applications of machine learning in finance 

    • Predicting credit card fraud

    • Interpreting neural network decisions 

    • Trading US treasury bonds 

Terry Benzschawel

Founder and principal

Benzschawel Scientific

Terry Benzschawel recently started his own firm after 30 years as a quant on Wall Street. The firm specializes in financial education, advanced model development, and systematic trading. Before that, Terry was a Managing Director in Citigroup's Institutional Clients Business, heading the Quantitative Credit Trading group.
Terry received a 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, Ophthalmology, and engineering prior to embarking on a career in finance. Terry began his financial career in 1988 at Chase Manhattan Bank, building genetic algorithms to predict corporate bankruptcy. In 1990, he moved to Citibank and trained a neural network to detect fraud on credit card transactions. In 1992 he was hired by Salomon's Fixed Income Arbitrage Group to build models for proprietary fixed income trading. In 1998, he moved to Citi’s Fixed Income Strategy department as a credit strategist with a focus on client-oriented solutions across all credit markets where he worked in related roles since then. Terry is a frequent speaker at industry conferences and events and has lectured on credit modelling at major universities and government institutions. In addition, he has published over a dozen articles in refereed journals and has authored two books:  CREDIT MODELING: FACTS, THEORIES AND APPLICATIONS and CREDIT MODELING: ADVANCED TOPICS.

10:3010:45

Morning break

10:30 - 10:45

10:4512:00

Introduction to deep learning neural networks

10:45 - 12:00

  • Major types of deep learning networks 

  • Convolutional neural networks  

  • Recurrent neural networks  

  • Long short-term memory networks 

  • Financial applications of deep learning models 

    • Machine learning model for corporate bond relative value

    • Predicting market moves from trading patterns 

      • Applying machine learning to risk, pricing, and other models 

      • Natural language processing and sentiment analysis 

  • Closing thoughts 

    • Neural networks versus decision trees 

    • How AI/ML is changing financial markets

Terry Benzschawel

Founder and principal

Benzschawel Scientific

Terry Benzschawel recently started his own firm after 30 years as a quant on Wall Street. The firm specializes in financial education, advanced model development, and systematic trading. Before that, Terry was a Managing Director in Citigroup's Institutional Clients Business, heading the Quantitative Credit Trading group.
Terry received a 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, Ophthalmology, and engineering prior to embarking on a career in finance. Terry began his financial career in 1988 at Chase Manhattan Bank, building genetic algorithms to predict corporate bankruptcy. In 1990, he moved to Citibank and trained a neural network to detect fraud on credit card transactions. In 1992 he was hired by Salomon's Fixed Income Arbitrage Group to build models for proprietary fixed income trading. In 1998, he moved to Citi’s Fixed Income Strategy department as a credit strategist with a focus on client-oriented solutions across all credit markets where he worked in related roles since then. Terry is a frequent speaker at industry conferences and events and has lectured on credit modelling at major universities and government institutions. In addition, he has published over a dozen articles in refereed journals and has authored two books:  CREDIT MODELING: FACTS, THEORIES AND APPLICATIONS and CREDIT MODELING: ADVANCED TOPICS.

12:0013:00

Lunch

12:00 - 13:00

13:0014:30

Machine learning for long-term investing

13:00 - 14:30

  • Overcoming a rapidly changing global economy

  • Using reinforcement learning over deep learning

  • Avoiding the endless amount of economic noise

  • How to make sense of so many fundamental data sets?

  • Which machine learning to use?

  • Why is HFT so much easier for machine learning than long-term?

Alexander Fleiss

CEO

RebellionResearch.com

Alexander Fleiss serves as CEO of RebellionResearch.com an online financial advisory & hedge fund that invests across all asset classes and utilizes a proprietary Machine Learning that monitors data from 53 countries on a daily basis. Mr. Fleiss has spoken about Artificial Intelligence investing in the Wall Street Journal, Fox News, BusinessWeek, Bloomberg News, MIT Technology Review, Wired, Mathematical Association of America, Financial Times, CNBC, Geo Magazine, Institutional Investor and the Wall Street Journal Reporter Scott Patterson’s book Dark Pools. In addition, Mr. Fleiss has lectured on Artificial Intelligence & Machine Learning at Princeton University, Amherst College, Yale School of Management, Booth School of Business at the University of Chicago, Tufts University, Cornell University, The Wharton School of Business at The University of Pennsylvania and Columbia Business School.

Prior to co-founding RebellionResearch.com in 2007, Mr. Fleiss served as a Principal at KMF Partners LP, a long-short US equity fund. Mr. Fleiss began his investment career as an analyst for Sloate, Weisman, Murray & Co which was acquired by Neuberger Berman. Mr. Fleiss developed investment algorithms with the firm’s CEO, Laura Sloate who is now a partner at Neuberger Berman and is one of the investors featured in Peter Tanous’ book Investment Gurus. Mr. Fleiss received a BA Degree from Amherst College.

14:3014:45

Afternoon break

14:30 - 14:45

14:4516:15

Big data's dirty secret

14:45 - 16:15

  • Impact of data quality on data analysis

  • Issues in assessing data quality in the big data space

  • How to clean large quantities of data

Harvey Stein

Head of the quantitative risk analytics group

Bloomberg

Dr. Harvey J. Stein is Head of the Quantitative Risk Analytics Group at Bloomberg, responsible for all quantitative aspects of Bloomberg's risk analysis products. Dr. Stein is well known in the industry, having published and lectured on mortgage backed security valuation, CVA calculations, interest rate and FX modeling, credit exposure calculations, financial regulation, and other subjects. Dr. Stein is also on the board of directors of the IAQF, an adjunct professor at Columbia University, a board member of the Rutgers University Mathematical Finance program and of the NYU Enterprise Learning program, and organizer of the IAQF/Thalesians financial seminar series. He received his BA in mathematics from WPI in 1982 and his PhD in mathematics from UC Berkeley in 1991.

16:1516:15

End of day one

16:15 - 16:16

Day two: Wednesday, March 18, 2020

08:3009:00

Registration and refreshments

08:30 - 09:00

09:0010:30

A deeper dive on natural language processing

09:00 - 10:30

  • Reading the news quickly to get in front of the crowd

  • What are your competitors doing?

  • Identifying other players in markets

  • Assessing the reaction to events, products, and services

  • Applying language models to analyze non-linguistic data

Elliot Noma

Machine Learning Engineer Consultant

Federal Reserve Bank of New York​​​​​​​

Elliot Noma has worked in both finance and data science. He is a consultant on machine learning and natural language processing solutions at the Federal Reserve Bank of New York. He also runs a proprietary trading company, Garrett Asset Management.

Dr. Noma taught machine learning at Columbia University and currently teaches quantitative risk management in the Masters in Mathematical Finance program at Rutgers University. 

Previously, Dr. Noma was a portfolio manager running a fund of hedge funds and was the Chief Risk Officer at Asset Alliance, a $3 billion seeder of hedge funds. 

Prior to working at Asset Alliance, Elliot was a risk manager at both Merrill Lynch Investment Advisers and Deutsche Bank, was an analyst at both JP Morgan and Salomon Brother, and was a investment banker at Chase. 

He has patented a levered ETF structure, has written a book on psychometrics, and has published scholarly articles in behavioral finance, psychology, information science, sociology, math, and finance. His most recent article in 2018 describes a new method of measuring financial risk aversion using hypothetical investment preferences.

Dr. Noma has a Ph.D. in the mathematical modeling of psychological processes from the University of Michigan along with an M.A. in Mathematics. He graduated from Dartmouth College with a B.A. in Mathematics.


10:3010:45

Morning break

10:30 - 10:45

10:4512:00

Deep learning for sequences and timeseries

10:45 - 12:00

  • Sequence data is a rich source of insight: timeseries of prices, text (sequence of words), video (sequence of images)

  • Sequences pose added complexity compared to non-sequence data

    • Order is important

    • Length of sequence is variable, not fixed

  • Deep dive into various methods, of increasing power, for dealing with sequences

  • Convolutional neural networks

  • Recurrent architectures: RNN, LSTM, GRU

  • Transformers

  • Examining important concepts like residual connections, attention, and self-attention

Ken Perry

Professor

NYU Tandon School of Engineering

Ken created the Risk Management department at Och Ziff and served as Chief Risk Officer for over 13 years. He led the Firm through a five-fold increase in AUM and headcount, the transition from private to public company, and managed major and minor financial/business crises and the introduction of new strategies and products.
He pioneered the use of quantitative techniques for portfolio construction and analysis at a fundamentally oriented firm, anticipating the “quantamental” revolution. Most recent activities have focused on how Artificial Intelligence may adapted to Finance.

12:0013:00

Lunch

12:00 - 13:00

13:0014:30

Speeding up option pricing with VDR

13:00 - 14:30

  • Discuss issues with option pricing speed

  • Present existing approaches, including machine learning approaches

  • Decompose the problem into market behavior and pricer behavior

  • Present the virtual dimensional reduction approach, which leverages this analysis

Harvey Stein

Head of the quantitative risk analytics group

Bloomberg

Dr. Harvey J. Stein is Head of the Quantitative Risk Analytics Group at Bloomberg, responsible for all quantitative aspects of Bloomberg's risk analysis products. Dr. Stein is well known in the industry, having published and lectured on mortgage backed security valuation, CVA calculations, interest rate and FX modeling, credit exposure calculations, financial regulation, and other subjects. Dr. Stein is also on the board of directors of the IAQF, an adjunct professor at Columbia University, a board member of the Rutgers University Mathematical Finance program and of the NYU Enterprise Learning program, and organizer of the IAQF/Thalesians financial seminar series. He received his BA in mathematics from WPI in 1982 and his PhD in mathematics from UC Berkeley in 1991.

14:3014:45

Afternoon break

14:30 - 14:45

14:4516:15

Machine learning and robust portfolio construction - is your investment manager artificially intelligent

14:45 - 16:15

  • From artificial intelligence to quantitative investing

  • The importance of purifying data

  • Robust portfolio construction with multiple factor risk budgeting

  • Quantamental and ESG

Patrick Dugnolle

U.S. Head of Multi-Assets and Quantitative Solutions

BNP Paribas Asset Management

Patrick Dugnolle, PhD, U.S. Head of Multi-Asset and Quantitative Solutions

Patrick is the U.S. Head of Multi-Asset and Quantitative Solutions at BNP Paribas Asset Management. In this role he is responsible for developing the strategy and products for our Multi-Asset and Quantitative Solutions business in the U.S. He is based in the firm’s New York office.

He joined BNP Paribas more than 16 years ago and his contribution to quantitative factor investing dates back to 2004 when he joined the Risk-Managed-Funds Research Team of CooperNeff Advisors in King of Prussia, PA. Patrick moved back to Paris in 2006 in order to manage a $100 million long-short portfolio of U.S. equities, and later he became Head of Financial Engineering for Fixed Income for BNP Paribas Asset Management. Since 2013, Patrick has largely contributed to the reengineering of the Multi-Factor Quantitative Equity Investment process, which currently has over $2 billion under management. Prior to moving to New York City in September 2018, he was managing more than $1.5 billion of assets invested in U.S., European and global equities.

Patrick holds a PhD in Computer Science applied to Theoretical Biology from the University Joseph Fourier of Grenoble and an Engineer Degree in Electronics from ICPI Lyon.
 

16:1516:15

End of course

16:15 - 16:16