Physics-Driven ML-Based Modelling for Correcting Inverse Estimation
This addresses the critical need to prevent disastrous consequences from failed estimations in domains like aero engine design, though it is incremental as it builds on existing optimization and machine learning techniques.
The paper tackles the problem of detecting and correcting failed state estimations in science and engineering inverse problems by proposing GEESE, a method that flags estimations with high physical model errors and corrects them via optimization, resulting in fewer failures and reduced need for physical evaluations compared to state-of-the-art approaches.
When deploying machine learning estimators in science and engineering (SAE) domains, it is critical to avoid failed estimations that can have disastrous consequences, e.g., in aero engine design. This work focuses on detecting and correcting failed state estimations before adopting them in SAE inverse problems, by utilizing simulations and performance metrics guided by physical laws. We suggest to flag a machine learning estimation when its physical model error exceeds a feasible threshold, and propose a novel approach, GEESE, to correct it through optimization, aiming at delivering both low error and high efficiency. The key designs of GEESE include (1) a hybrid surrogate error model to provide fast error estimations to reduce simulation cost and to enable gradient based backpropagation of error feedback, and (2) two generative models to approximate the probability distributions of the candidate states for simulating the exploitation and exploration behaviours. All three models are constructed as neural networks. GEESE is tested on three real-world SAE inverse problems and compared to a number of state-of-the-art optimization/search approaches. Results show that it fails the least number of times in terms of finding a feasible state correction, and requires physical evaluations less frequently in general.