Friday April 5th 2019
Director, Portfolio Analytics Bank of America Merrill Lynch
Cristian Homescu: Director, Portfolio Analytics Bank of America Merrill Lynch
Cristian is part of the Portfolio Analytics team within Chief Investment Office, Global Wealth and Investment Management division Bank of America Merrill Lynch. He is developing and investigating quantitative solutions in areas such as investment strategies, goals-based wealth management, asset allocation, machine learning and big data analysis, factor-based investing and risk factor models, portfolio risk and attribution, stress testing and scenario construction. He is very interested in application of state-of-the-art algorithms and numerical methods in wealth and investment management, and in high-performance computing. Prior to joining Bank of America Merrill Lynch, Cristian was a front office quant for Wachovia and Wells Fargo. After supporting interest rate trading desk, he was the lead quant for FX and Commodities trading desks. He has a PhD from Florida State University in computational and applied mathematics, and MSc degrees from University of Paris XI and University of Craiova.
- This session will analyse the emerging techniques applicable to quantum computing and its applications.
Co-Founder, New York Quantum Computing Meet-up & XVA Quant Core Lead, Wells Fargo
Steve Yalovitser: Co-Founder, New York Quantum Computing Meet-up and Director, XVA Quant Core Lead, Wells Fargo
- This presentation will discuss how machine learning algorithms can be used to study and evaluate market microstructure.
Quantitative Researcher, Guggenheim Partners
Zhibai Zhang: Quantitative Researcher, Guggenheim Partners
Non-negative matrix factorization (NMF) is a widely-used tool for analysing high-dimensional datasets. Its popularity stems from its ability to extract meaningful factors from the data. Applications include image processing, text mining and bioinformatics. In this talk we will give an overview of NMF and demonstrate our implementations of recent NMF algorithms by automatically classifying a series of websites based on their content. We will then briefly discuss applications of NMF in finance.
Technical Consultant, NAG
Edvin Hopkins: Technical Consultant, NAG
Edvin first worked with NAG between 2010 and 2013, as part of a Knowledge Transfer Partnership with the University of Manchester. Long-time NAG collaborator Professor Nick Higham and his team had developed many new algorithms to compute matrix functions. Edvin’s role was to convert these algorithms into code for the NAG Library.
After the successful collaboration, Edvin worked with Professor Higham as a post-doctoral research associate, before finally joining NAG in 2015. He is based in our Manchester Office.
Edvin gained a PhD in Numerical Relativity from the University of Cambridge in 2009. His supervisor was Dr John Stewart. This followed a first class honours degree in Mathematics and a “Certificate of Advanced Study in Mathematics” from the same institution.
- Data prep and feature engineering: Is the AI built over the data based on solid foundations?
- Issues of overfitting and maximizing the signal to noise ratio
- Evaluating your algorithm choice: what do you want to achieve?
- Understanding fake signals: when machine learning fails
Abstract: To gain an edge in the markets quantitative hedge fund managers require automated processing to quickly extract actionable information from unstructured and increasingly non-traditional sources of data. The nature of these “alternative data” sources presents challenges that are comfortably addressed through machine learning techniques. We illustrate use of AI and ML techniques that help extract derived signals that have significant alpha or risk premium and lead to profitable trading strategies.
This session will cover the following topics:
- The broad application of machine learning in finance
- Extracting sentiment from textual data such as news stories and social media content using machine learning algorithms
- Construction of scoring models and factors from complex data sets such as supply chain graph, options (implied volatility skew, term structure) and ESG (Environmental, Social and Governance)
- Robust portfolio construction using multi-factor models by blending in factors derived from alternative data with the traditional factors such as fama-french five-factor model.
Quantitative Research Solutions, Bloomberg, LP
Arun Verma: Quantitative Research Solutions, Bloomberg, LP
Dr. Arun Verma joined the Bloomberg Quantitative Research group in 2003. Prior to that, he earned his Ph.D from Cornell University in the areas of computer science & applied mathematics. At Bloomberg, Mr. Verma’s work initially focused on Stochastic Volatility Models for Derivatives & Exotics pricing and hedging. More recently, he has enjoyed working at the intersection of diverse areas such as data science (for structured & unstructured data), innovative quantitative & machine learning methods and finally interactive visualizations to help reveal embedded signals in financial data.