As Ben Dickson outlines, Geoffrey Hinton has introduced a new type of neural network algorithm, the 'forward-forward algorithm'. "The idea behind the forward-forward algorithm is to replace the forward and backward passes of backpropagation with two forward passes. The two passes are similar, but they work on different data and have opposite objectives." The new algorithm can work in cases where back-propagation algorithms can't, and more closely models actual human neural networks. "Connections between different cortical areas do not mirror the bottom-up connections of backpropagation-based deep learning models. Instead, they go in loops, in which neural signals traverse several cortical layers and then return to earlier areas." Here's Hinton's paper.
Today: 0 Total: 18 [Direct link] [Share]
