Physical Review Journal Club: Machine Learning in Fluid Dynamics: A Critical Assessment
Tuesday, November 4, 2025 at 5:30 p.m. ET
Video Link
The Physical Review Journal Club is back with a discussion with author Kunihiko “Sam” Taira (UCLA), presenting their recently published Physical Review Fluids perspective article “Machine Learning in Fluid Dynamics: A Critical Assessment.” After the presentation, the authors will answer attendee questions in a live Q&A session, moderated by PRFluids Associate Editor Steven Brunton (University of Washington).
Summary:
In recent years, machine learning (ML) has become increasingly popular in the fluid dynamics community, offering new tools for analyzing, modeling, predicting, and controlling a wide range of flows. In their perspective article, Kunihiko “Sam” Taira and co-authors provide a critical assessment of both the opportunities and limitations of ML. While machine learning has in some cases outperformed traditional approaches, many fundamental challenges remain; tackling them is necessary both to deepen our understanding of flow physics, and to extend the applicability of ML beyond fundamental research.
To accelerate progress in the field, the authors highlight the importance of community-maintained datasets and open-source code repositories, as well as the need for effective training, both for early-career and well-established researchers. This article aims to spark discussions and foster collaborative efforts toward a more robust integration of machine learning into fluid dynamics research.
When: Tuesday, Nov. 4, 2025 — 5:30 p.m. Eastern Time / 2:30 p.m. Los Angeles / 10:30 p.m. London / 7:30 a.m. Sendai (next day, Nov. 5)
Read ahead:
Machine Learning in Fluid Dynamics: A Critical Assessment
Kunihiko Taira, Georgios Rigas, Kai Fukami
Phys. Rev. Fluids 10, 090701
