Applications of AI for quantitative finance at Thomson Reuters

Description

In March 2018, a financial investment AI startup announced that it was being acquired for $550 million—the largest AI acquisition to date. While the size of the deal was surprising in itself, many were even more surprised that it happened in the financial investment domain, which has often lagged other industries in applying nontraditional techniques. In the last couple of years, the finance industry has come a long way toward embracing new techniques associated with artificial intelligence, a fact due to both technology advances and a growing understanding of and comfort with these techniques by the financial community.

The Thomson Reuters Lab in San Francisco—a team of data scientists and researchers creating quantitative models for investors—is exploring novel uses of alternative data, new machine learning algorithms to generate alpha and measure risk, and how to improve the signal-to-noise ratio for investors. Some of these research projects and the creation of new derived data wouldn’t have been possible a few years ago: extracting meaning from unstructured text, identifying objects in images, and understanding the relationships between events and companies in a complex network all require a combination of recent advances in cloud computation, algorithms, and data access. Joe Rothermich explores some of the lab’s projects, including credit risk analysis, the creation and use of alternative data, and news event detection, and explains how advances in AI made these projects possible. Joe also shares what the lab is learning along the way and the challenges of applying these technologies to investment models.