Title: Model-Free Forecasting with Applications to Multi-Sensor Arrays
Speaker: Tyrus Berry, George Mason University
Date: March 16, 2021
Summary:
The talk by Tyrus Berry at the ACMD Seminar focused on model-free forecasting methods, particularly their application to multi-sensor arrays. Berry discussed the key considerations in choosing machine learning methods for this purpose, including the basis, regularizer, and optimization scheme.
The Diffusion Maps algorithm was highlighted as an effective method for approximating the optimal basis for a wide range of stochastic systems on manifolds. Berry also explored ways to combine model-free methods with imperfect models to mitigate model errors.
Potential applications of these techniques include using multi-sensor arrays for calibration and drift compensation in various systems.
Tyrus Berry is an assistant professor at George Mason University, continuing his research on manifold learning with applications to dynamical systems.
Source: https://www.nist.gov/itl/math/acmd-seminar-model-free-forecasting-applications-multi-sensor-arrays
Keywords: Quantum, Stochastic, Forecast
