6 March 2018
Present the foundations of machine learning and then introduce the tools for building and training deep neural networks. Finally, provide example of applications in multiple domains; e.g. natural language processing and Image classification.
This is a beginner’s course that firstly introduces the foundations of deep learning, the theory and concepts behind it all. Then a practical session is offered to attendees to implement neural networks and apply them to industrial use cases.
Pre-requisites and technical information:
♦ The programming language used will be Python so basic knowledge of Python is desirable, but not essential. Experience with alternative programming languages is also helpful.
♦ The tools used in this workshop are all open source. They are TensorFlow, Keras and Python 3.6.
Presentation 1: Foundations of Deep Learning
Presenter: Barbara Fusinska, Data Scientist
Barbara Fusinska is a Data Scientist with strong software development background and over than thirteen years of both commercial and academic experiences. Work on her master’s degree dissertation and research activities have brought her focus to the Data Science field.
She believes in the importance of data and metrics when growing a successful business and the significance that Machine Learning and Artificial Intelligence bring when gaining insights and drawing conclusions. She is passionate about exploring, continually learning and sharing her knowledge via writing and speaking at conferences with the aim of helping others understand the Data Science area.
Presentation 2: Applications of Deep Learning: Explained with Use Cases
Presenter: Jayadeep Shitole, OptiRisk Systems
In recent times, the explosive growth of data and its harnessing through Artificial Intelligence have impacted most of the B2B and B2C businesses and markets. Naturally there is considerable interest in acquiring knowledge and skills in this field. In this half day tutorial, we introduce the motivations, concepts, models and algorithms which are used in deep learning. We then explain how it is applied in domains of natural language processing and image classification. The illustrative prototype applications are constructed using TensorFlow. Participants will learn how to build and deploy deep learning models using TensorFlow.
Jayadeep Shitole is a Research Analyst and Software Developer at OptiRisk Systems. He provides support to hedge fund clients of OptiRisk Systems in setting up their algorithmic trading systems. He has also worked as a Data Scientist for a leading Big Data analytics company where he worked on designing, developing, and deploying data-driven predictive models to solve business problems using machine learning and statistical modelling.
8 March 2018
The workshop has 3 different sessions. Sessions 1 and 2 by Ernie Chan and Humberto Brandao focus on the usage of Machine Learning for building Algorithmic Trading strategies. The 3rd session by Rajib Ranjan Borah explores the various components of Algorithmic Trading Systems, the technological complexities in these systems and at the exchanges (and opportunities thereof), and discusses ways to increase profitability through proper system design.
Presenter: Ernie Chan, Hedge Fund Manager, QTS
Avoiding Overfitting in Machine Learning
Insufficient historical data is a major hurdle in building a trading model free from data snooping bias. My talk will discuss several techniques, some borrowed from machine learning, that can alleviate overfitting and enhance the statistical significance of a backtest.
Ernie Chan is the Managing Member of QTS Capital Management, LLC., a commodity pool operator and trading advisor. Ernie has worked for various investment banks (Morgan Stanley, Credit Suisse, Maple) and hedge funds (Mapleridge, Millennium Partners, MANE) since 1997. He received his Ph.D. in physics from Cornell University and was a member of IBM’s Human Language Technologies group before joining the financial industry. He is the author of “Quantitative Trading: How to Build Your Own Algorithmic Trading Business”, “Algorithmic Trading: Winning Strategies and Their Rationale”, and “Machine Trading: Deploying Computer Algorithms to Conquer the Markets”. Find out more about Ernie at www.epchan.com.
Presenter: Rajib Ranjan Borah, CEO iRage
Different Components of Algorithmic Trading Systems – improving trading profitability by designing optimal trading systems.
By properly leveraging the power of technology, a trader can increase the profitability of an already profitable systematic trading strategy multi-fold. This talk looks at the evolution of algorithmic trading systems – and the efficiency introduced at each step. We also introduce the participants to the various technological complexities at exchanges – and opportunities that could exist because of the same. The presentation format encourages the participants to have a discussion which is open-ended, and break down technical complexities to functional implications in a way that will be of value to quantitative traders. The content is deliberately presented in a qualitative form, not too technical, and explains complex topics from a functional perspective using anecdotes and examples.
Rajib is the co-founder & CEO of iRage, one of India’s leading High-Frequency Trading firms, which manages potentially the broadest option portfolio book in India. He is also the co-founder and director of QuantInsti, an ‘Algorithmic and Quantitative Trading’ training and research institute which has trained thousands of professionals from over 130 countries. His prior experiences include high-frequency trading on all major US & European exchanges (Optiver, Amsterdam); data analytics technology (Oracle); business strategy for a trading firm & derivatives exchanges (Strategy Consulting, PwC). Rajib has thrice represented India at the World Puzzle Championship. He was also a finalist at the Indian National Biology Olympiad (top 24 nationwide). Rajib holds an MBA from IIM Calcutta, a bachelor’s degree in Computer Engineering from NIT Surathkal; and has internship experiences with Bloomberg in New York (derivatives research) & Solutia’s EMEA strategy HQ in Belgium.
Presenter: Humberto Brandão, Professor and Head Scientist of R&D Lab, Universidade Federal de Minas Gerais
Applying Machine Learning in different kinds of algorithmic trading strategies
The objective of this course is to show you how to create databases from your own strategies and adapt it for Machine Learning Methods. Besides presenting different generic algorithmic trading strategies, some machine learning methods are also explained with a discussion about different kinds of validation processes.
Humberto Brandão is the Head of the Research & Development Lab (R&D Lab) at Federal University of Alfenas (Brazil), where he is also a Professor. He has been working on Algorithmic Trading using Machine Learning since 2009. During this period, he created a realistic simulator, which has been used for High-Frequency Trading in Brazil. As a consultant for hedge funds, Humberto has been applying different techniques in order to improve their return and risk over different kind of strategies. Recently, Humberto won several important prizes in competitions related to Algorithmic Trading and Data Science.