Gradients without Backpropagation
This addresses the computational bottleneck of backpropagation in machine learning, offering a novel approach for faster gradient descent, though it may be incremental in its application to existing methods.
The paper tackles the problem of computing gradients for optimization without backpropagation by introducing the forward gradient method, which uses directional derivatives from forward-mode automatic differentiation, resulting in training up to twice as fast in some cases.
Using backpropagation to compute gradients of objective functions for optimization has remained a mainstay of machine learning. Backpropagation, or reverse-mode differentiation, is a special case within the general family of automatic differentiation algorithms that also includes the forward mode. We present a method to compute gradients based solely on the directional derivative that one can compute exactly and efficiently via the forward mode. We call this formulation the forward gradient, an unbiased estimate of the gradient that can be evaluated in a single forward run of the function, entirely eliminating the need for backpropagation in gradient descent. We demonstrate forward gradient descent in a range of problems, showing substantial savings in computation and enabling training up to twice as fast in some cases.