Benchmarking Decoupled Neural Interfaces with Synthetic Gradients
This addresses the need for parallelization in large-scale machine learning by enabling asynchronous updates, though it appears incremental as it builds on existing synthetic gradient concepts.
The paper tackled the problem of strict coupling in artificial neural networks that prevents asynchronous weight updates, by benchmarking synthetic gradients with decoupled neural interfaces (SG-DNI) against standard methods. The result showed that SG-DNI not only captures the learning problem but is also over 3-fold faster due to its asynchronous capabilities.
Artifical Neural Networks are a particular class of learning systems modeled after biological neural functions with an interesting penchant for Hebbian learning, that is "neurons that wire together, fire together". However, unlike their natural counterparts, artificial neural networks have a close and stringent coupling between the modules of neurons in the network. This coupling or locking imposes upon the network a strict and inflexible structure that prevent layers in the network from updating their weights until a full feed-forward and backward pass has occurred. Such a constraint though may have sufficed for a while, is now no longer feasible in the era of very-large-scale machine learning, coupled with the increased desire for parallelization of the learning process across multiple computing infrastructures. To solve this problem, synthetic gradients (SG) with decoupled neural interfaces (DNI) are introduced as a viable alternative to the backpropagation algorithm. This paper performs a speed benchmark to compare the speed and accuracy capabilities of SG-DNI as opposed to a standard neural interface using multilayer perceptron MLP. SG-DNI shows good promise, in that it not only captures the learning problem, it is also over 3-fold faster due to it asynchronous learning capabilities.