A survey of deep learning optimizers -- first and second order methods
This is a survey paper that summarizes existing methods for researchers and practitioners in deep learning, making it incremental in nature.
The paper provides a comprehensive review of 14 standard optimization methods used in deep learning, addressing challenges like saddle points and ill-conditioning, but does not present new experimental results or concrete performance numbers.
Deep Learning optimization involves minimizing a high-dimensional loss function in the weight space which is often perceived as difficult due to its inherent difficulties such as saddle points, local minima, ill-conditioning of the Hessian and limited compute resources. In this paper, we provide a comprehensive review of $14$ standard optimization methods successfully used in deep learning research and a theoretical assessment of the difficulties in numerical optimization from the optimization literature.