Presented By
O’Reilly + Intel AI
Put AI to Work
April 15-18, 2019
New York, NY
Discover opportunities for applied AI
Organizations that successfully apply AI innovate and compete more effectively. How is AI transforming your business?
Be a part of the program—apply to speak by October 16.

GAIA: The Global AI Allocator

Aric Whitewood (WilmotML)
4:05pm4:45pm Thursday, April 18, 2019
Implementing AI
Location: Regent Parlor
Secondary topics:  AI case studies, Financial Services, Temporal data and time-series

Who is this presentation for?

Data Scientist, CEO, COO, Quantitative Researcher, Portfolio Manager, Investor, Economist

Level

Beginner

Prerequisite knowledge

This presentation does not require specialist knowledge of the domain - finance / investment management, but does require some basic knowledge of machine learning. Those members of the audience with experience in both areas will benefit more. Some technical information will be included, but the emphasis is more on the application of AI rather than detailed mathematical expositions for example. The intent is to make the subject matter approachable, to discuss some of the key changes happening in investment management, and to explain in clear terms our own unique AI approach within this fascinating domain.

What you'll learn

This presentation is concerned with the application of AI to investment management and finance in general. More specifically, the key ideas include: how the world of investment management is changing as a result of this new technology, transparency and 'explainability' in financial data science (we have designed our system to be a glass box rather than a black box), the combination of human and machine intelligence (we believe it is this combination which will provide the most benefit from AI technologies in the coming years), the use of custom hardware for machine learning (we are using FPGAs for our algorithms), and the development of more timely, alternative data derived macro indicators to help investment decisions. These key ideas should be interesting and useful to both practitioners/specialists and non-specialist interested parties.

Description

We are a boutique asset management firm, specialising in macro and machine learning. The two founders previously worked in Credit Suisse as the Chief Strategist and Head of Data Science, respectively. Our firms philosophy on AI is that human and machine combined is more powerful than machine on its own, and providing both transparency and trust in AI systems is essential in the financial services domain. This presentation details our research on AI, including the design and development of our prediction engine, called GAIA (the Global AI Allocator), which has been running in production since January 2018. Some specific topics include: how to combine fundamental and quantitative skills effectively (a challenge for many incumbent firms), how to approach the problem of generating meaningful predictions with highly noisy, non-stationary data sets (financial time series), transparency and ‘explainability’ in the context of investment management (why this is important from an investor and regulatory perspective, and how we achieve it), and the use of heterogeneous compute architectures, specifically GPUs and FPGAs. In addition, the topic of creating macro relevant indicators from alternative data (newspapers and social media) will also be covered, such as real time unemployment and geopolitical risk indicators.

Photo of Aric Whitewood

Aric Whitewood

WilmotML

Aric Whitewood is co-founder of WilmotML, a machine learning and macroeconomics focused investment and advisory firm. He is also an Honorary Senior Lecturer in the Computer Science Department of University College London (UCL), for which he runs several research programs with UCL students on machine learning topics.

He focuses on the combination of neuroscience, artificial intelligence (A.I.), and investing, with a particular emphasis on developing investment systems that are transparent (enabling trust in investment decisions) and that operate on longer timescales than has historically been the case with algorithmic systems (typically months).

Previous to his current position, he was Head of Data Science in Credit Suisse Zurich, where he ran A.I. projects across a number of businesses and geographic locations. He also served as the Banks subject matter expert in machine learning, regularly presenting to both the Banks management as well as its major clients.

He holds a PhD in Electronic Engineering from UCL (2006).

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)