Data Assimilation

Parameterized VAE-transformer ROM diagram

Efficient adaptation of ROMs for unsteady flows using data assimilation

This paper proposes an efficient retraining strategy for parameterized reduced-order models, which adapts a variational autoencoder and transformer architecture to out-of-sample flow regimes from sparse observations.

Real-time forecasting of chaotic dynamics from sparse data and autoencoders

This paper introduces DA-MIRL, a novel framework that integrates sequential DA with off-policy RL to estimate and control high-dimensional, spatio-temporal chaos using only partial and noisy observations..

Data-assimilated model-informed reinforcement learning (DA-MIRL)

This paper introduces DA-MIRL, a novel framework that integrates sequential DA with off-policy RL to estimate and control high-dimensional, spatio-temporal chaos using only partial and noisy observations.

Real-time digital twins: unifying bias-aware data assimilation and machine learning

Invited seminar at Data-Centric Engineering Journal (Cassyni).

Real-time data assimilation in nonlinear thermoacoustic instabilities

Invited talk at the Phi-ML meets Engineering seminar series, The Alan Turing Institute.

A real-time digital twin of azimuthal thermoacoustic instabilities

This paper develops novel methods to infer unknown unknowns in complex dynamical systems using data assimilation techniques.

Inferring unknown unknowns

This paper develops novel methods to infer unknown unknowns in complex dynamical systems using data assimilation techniques.