Big Data & Finance: An Interview With Veteran Quant Steve Cannon
“AI” has become one of the biggest buzzwords in the technology sector. Artificial Intelligence (AI) allows computer systems to perform tasks that normally require human intelligence ‒ tasks such as visual perception, decision-making, and translating between languages.
Some examples of AI technology are machine learning (ML) and natural language processing (NLP). ML allows computer algorithms to improve their own functionality by analyzing data and interpreting “feedback” from their own functions and experiences.
NLP technology allows computers to translate natural language (the language that humans read and write in) into a format that a computer program can interpret and use. Claims of AI systems “revolutionizing” the world have pervaded online media outlets ‒ but what impact will AI technology really have on global markets and lifestyles?
To learn more about this, we sat down with Steve Cannon to discuss his time as the Head of Model Data at AQR Capital & Head of Big Data at Two Sigma.
Steve Cannon grew up near Park City, Utah, and has a passion for German automotive engineering, the Italian opera, and minimalist art. Cannon studied undergraduate Physics at Columbia University, although his interests began to shift towards finance during his senior year of college, when he received a letter from David Shaw inviting him to interview at D.E. Shaw, his esteemed global investment firm.
Cannon ultimately decided to pursue graduate Physics at The University of Texas at Austin, but Shaw’s letter left a lasting impression: “I still have that amazing letter.” Cannon’s interest in data began at Interactive Brokers, who hired Cannon as a data engineer: “I thought, ‘OK, let’s learn data then.’ I had no idea how much that decision would shape my career. I discovered my passion.” Cannon notes that studying Physics has left him with a valuable lesson throughout his career path: “Just keep trying. I’ve taken physics tests in classes full of geniuses where the average was below 50%... You all come in the next day and try again, a little harder.”
AQR Capital is currently working to find ways to apply ML to optimize its investment decisions and portfolios. When we asked Cannon about the current use of ML and NLP in the financial sector, he had much to say about their current limitations and future potential.
While Cannon acknowledged that ML can be a tremendous tool in predicting market behavior, he also noted that the usefulness of ML heavily relies on “the skill of the user,” so it’s unlikely that ML models will ever be able to function effectively without human involvement.
Moreover, Cannon emphasized that ML models are useful only when they can be understood: ML models typically answer “a game theoretical question,” so if a team of analysts don’t fully understand “the game,” then the associated model will not be of any use to them.
When asked about NLP, Cannon stressed that the technology is still very much in its “early days.” To demonstrate NLP’s current limitations, Cannon provided the example of search engines, which utilize NLP to translate user inputs into search results.
Cannon shared a screenshot of the search results on Amazon and Google that appear after entering “shirt without stripes:” the search engines erroneously return a large proportion of images of shirts with stripes, instead of shirts without stripes.
Cannon mentioned that it’s fairly difficult to predict the future impacts of AI in the financial sector: “In my experience, the people who know the most about what’s next in finance are likely to say nothing…” However, while the future can be hard to predict, Cannon advises students to study subjects in engineering, economics, the sciences, and math, as these disciplines often provide opportunities to use and develop the technologies that will ultimately improve global productivity and welfare.
To wrap up, Cannon offered two pieces of advice to the future generation of AI and data analysis: “be good to your family,” and “be unrealistically optimistic about the future.”
Written by Alex Ristic & Thomas Braun, Edited by Alexander Fleiss