2026/01/15 – In-Person and On-Line
LQG 2026/01/15 – 18:30 – Limits to (Machine) Learning – Semyon Malamud – IP *OLO* Hybrid
Seminar by Semyon Malamud
The LQG is very grateful to our hosts of this seminar: Snowflake
Limits to (Machine) Learning
Machine learning (ML) models are highly flexible but face fundamental limits when learning from finite data. We show that a universal lower bound—the Limits-to-Learning Gap (LLG)—captures the unavoidable gap between the true (population) benchmark and linear ML models. A researcher willing to measure the true population $R^2$ needs to adjust the model performance by this gap. In financial data, the estimated gaps are large, implying that standard ML methods may substantially understate predictability.
About the Speaker
Semyon Malamud
Associate Professor of Finance at the Swiss Federal Institute of Technology in Lausanne
and the Director of the Financial Engineering Section
Visiting Lecturer at Imperial College London
Semyon Malamud is an Associate Professor of Finance at the Swiss Federal Institute of Technology in Lausanne and the Director of the Financial Engineering Section.
He has a PhD in Mathematics from ETHZ, holds a Senior chair at the Swiss Finance Institute, is a Lamfalussy fellow of the European Central Bank, and a Research Fellow of the Centre of Economic Policy Research (CEPR) and the Bank for International Settlements. He is also an Associate Editor at the Journal of Finance.
Semyon’s research has been published in top economics and finance outlets, including Econometrica, American Economic Review, Journal of Finance, Review of Financial Studies, and the Journal of Financial Economics. His research has also been recognized with several awards, including a Jack Treynor Prize, two INQUIRE Europe prizes, the Dauphine-Amundi Chair in Asset Management award, the Europlace Institute of Finance award, and the ETF Academy Award.
