Neural Network Regularization via Robust Weight Factorization
This is an incremental improvement for deep learning practitioners, as it builds on existing dropout-like strategies to enhance regularization in neural networks.
The paper tackles the problem of overfitting in large neural networks by proposing a new regularization method called FaMe, which learns a robust factorization of weight matrices to improve generalization.
Regularization is essential when training large neural networks. As deep neural networks can be mathematically interpreted as universal function approximators, they are effective at memorizing sampling noise in the training data. This results in poor generalization to unseen data. Therefore, it is no surprise that a new regularization technique, Dropout, was partially responsible for the now-ubiquitous winning entry to ImageNet 2012 by the University of Toronto. Currently, Dropout (and related methods such as DropConnect) are the most effective means of regularizing large neural networks. These amount to efficiently visiting a large number of related models at training time, while aggregating them to a single predictor at test time. The proposed FaMe model aims to apply a similar strategy, yet learns a factorization of each weight matrix such that the factors are robust to noise.