The next appointment in the PhDTalks seminar series will take place on Wednesday, January 11th, 2026, in Grandori Room (Building 4), from 12:00 to 13:00 CET.

PhDTalks is a series of seminars and discussions among PhD candidates. The events aim to provide a space for networking among doctoral students and for engaging with the many projects developed within our department.

The speaker Teng Ma will deliver a seminar entitled “Learning governing equations of weakly nonlinear oscillators by machine learning”

At the end of the event, a light refreshment will be offered, sponsored by the department.

The seminar will also be accessible online at the following link.

Abstract
Weakly nonlinear oscillators are fundamental to a wide range of physical systems. While linear models are often used for simplicity, the long-term stability and evolution of these systems are frequently governed by subtle nonlinearities. In structural engineering and beyond, characterizing these weakly nonlinear signatures from data remains a significant challenge due to their small magnitude compared to the dominant linear response.
We introduce EvLOWN, a data-driven methodology designed to autonomously discover the governing equations of weakly nonlinear oscillators from sparse and noisy observations. After validating its robustness on benchmark systems, we demonstrate EvLOWN’s versatility through diverse high-impact applications: from reconstructing the orbital dynamics of the Tiangong and International Space Stations to capturing the precise equations of motion for complex vortex-induced vibrations (VIV) from wind-tunnel experimental data.

Speaker’s Bio
Teng is currently a double PhD candidate in Civil Engineering at Tongji University and Structural Seismic and Geotechnical Engineering at Politecnico di Milano. His research focuses on AI for Engineering Systems, with a specialization in data-driven dynamical system identification. He conceptualized and developed an interpretable scientific machine learning framework for governing-law discovery, with successful applications in fluid-structure interaction (FSI) and MEMS devices. Related researches are published on Structure and Infrastructure Engineering, Physics of Fluids, Journal of Wind Engineering and Industrial Aerodynamics etc.
In his spare time, Teng enjoys singing, boxing, running and hiking in the mountains.