Self-adaptive node-based PCA encodings
arXiv:1708.04498v1
AI Analysis
This is an incremental improvement for distributed computing and neural network efficiency.
The paper tackles the problem of distributed principal component analysis (PCA) by proposing Simple Hebbian PCA, which reduces the number of trainable weights by half by removing intralayer weights.
In this paper we propose an algorithm, Simple Hebbian PCA, and prove that it is able to calculate the principal component analysis (PCA) in a distributed fashion across nodes. It simplifies existing network structures by removing intralayer weights, essentially cutting the number of weights that need to be trained in half.