Xi Chen, who got his Ph.D. in machine learning from Carnegie Mellon University’s School of Computer Science, is an assistant professor of information, operations and management sciences at New York University’s Stern School of Business. He was recently named to the Forbes 30 Under 30 list. Chen featured in the science rankings, but he’s working on machine learning products that will disrupt the financial services industry.
How will machine learning impact financial services?
First of all, modern machine learning techniques will lead to more accurate predictions of the future prizes and trends in financial services. It can also help predict market trends by aggregating many opinions on the social networks. This might be even more accurate than the subjective judgment from an expert.
Also, it can lead to quick decisions.
For example, machine learning techniques have been widely used in algorithmic trading.
Moreover, using machine learning techniques, banks can know better about the needs of each individual and thus provide personalized financial services to each customer.
How are these leading to new learning methods and approaches to analyzing data?
Increasing demand for better financial services will also lead to improvement in learning methods and data analytics. For example, the flourishing of algorithmic trading will lead to new techniques, which can predict the prices on the millisecond and even microsecond level. It can also lead to cost-cutting.
Do you see a potentially negative impact of machine learning and automation on employment?
Yes. Machine learning technology will make many jobs obsolete – some will be replaced by machines in the future. Examples include driverless cars, automated trading systems, and robots as waiters and cleaning people, to name just a few. I think one way [to approach the problem of unemployment] is to provide people with a better education so that the next generation will be trained to perform jobs that machines are not competent in.
Generally, what interesting trends are you seeing in the areas your research focuses on?
I am working on several key aspects in machine learning and statistics, from data collection to structural data analysis.
For example, my series of work on crowdsourcing has laid the foundation for how to enable high-quality and cost-saving crowdsourcing services to harness the wisdom of the crowd to extract rich information from web-scale data. This line of research not only has an immense impact on utilizing human power for data-intensive scientific research, such as FoldIt, EyeWire and Zooniverse, but also helps boost the crowded labor markets and opens up opportunities for unemployed people.
In addition, I have been devoted to developing new methods in statistical machine learning to better analyze the high-dimensional data, which is prevailing in scientific studies. These methods have been successfully applied to analyzing climatological data, brain fMRI data and genetic data, leading to new scientific discoveries.
In addition, I am also investigating the application of machine learning to revenue management.
Crowdsourcing becomes one of the most important tools for collecting high-quality data. However, the reliability is always a key challenge in crowdsourcing due to the noise from workers. There is an increasing demand for new technologies to enable more reliable crowdsourcing systems for complex tasks.
For high-dimensional statistics, although this topic has been studied extensively in the statistical community over the past decade, the computational challenges of many developed models have not been fully explored. In my understanding, the research on new statistical models together with efficient implementation will be a growing trend.
Photo credit: NicoElNino/GettyImages