This Friday, we will host our third seminar speaker for the spring semester, Professor Dacheng Xiu, University of Chicago

Professor Xiu will discuss topics and papers related to, "Machine Learning in Asset Pricing: Myths, Criticism, and Real Limits" at 10:30 AM in the Finance Department conference room, 1200 RCB. 

Abstract:

"While machine learning has transformed empirical asset pricing—enabling high-dimensional modeling and reshaping the measurement of expected returns—it has also given rise to a number of misunderstandings and misplaced critiques. Common myths surrounding interpretability, robustness, profitability, and model comparison are examined, highlighting that many perceived shortcomings originate from the inherent characteristics of financial data rather than from the algorithms themselves. Particular emphasis is placed on the prevalence of weak signals and data scarcity, along with recent theoretical results that delineate when machine learning methods can—and cannot—effectively learn in high-dimensional settings. These insights reveal fundamental statistical limits of learning and underscore the importance of regularization, shrinkage, and theoretically informed model selection. Looking forward, deep learning, generative AI, and large language models offer new avenues for incorporating structure, priors, and alternative data into asset-pricing research. Collectively, these developments point toward a future in which machine learning complements, rather than replaces, financial economics by providing new tools, new insights, and a sharper understanding of the boundaries of what data alone can reveal."