Deep Learning From a Control Vantage Point: Feedforward, Residual, Transformers, Large Language Models, And Selective State-Space Models
This article provides a gentle introduction to recent advances in deep learning from a control theory perspective and discusses how control theory has recently been used to better understand and design deep neural networks.
Álvaro Rodríguez Abella, J. Silvestre, B. Gharesifard et al.
· IEEE Control Systems · 0 citations