Tensor Regression

Tensor regression focuses on modeling relationships between tensor-valued data—multi-dimensional arrays—and scalar or tensor responses, addressing the challenges posed by high-dimensional datasets common in fields like neuroimaging and materials science. Current research emphasizes developing efficient algorithms, such as tree-based methods and Bayesian approaches incorporating low-rank decompositions, to handle the computational complexity of these models, as well as addressing biases and ensuring interpretability. These advancements are significant for improving prediction accuracy and extracting meaningful insights from complex, high-dimensional data across diverse scientific disciplines and practical applications, including medical imaging analysis, process optimization, and human-robot interaction.