Target Optimization

Target optimization focuses on finding inputs that yield a desired output from a function, often a complex or stochastic "black box" model. Current research emphasizes developing robust methods for various model types, including Gaussian processes, neural networks (like LSTMs), and kernel methods, exploring optimal target variable selection and efficient surrogate function construction to reduce computational cost. This field is crucial for diverse applications, from controlling autonomous vehicles and optimizing industrial processes to improving machine learning algorithms by efficiently navigating complex loss landscapes.