Learning Representation from Neural Fisher Kernel with Low-rank Approximation
This provides a unified tool for representation extraction in both supervised and unsupervised learning, which is incremental as it builds on existing Fisher Kernel methods.
The paper tackles the problem of extracting high-quality representations from neural networks by defining the Neural Fisher Kernel (NFK) and showing it has low-rank structures, then proposes an efficient low-rank approximation algorithm that achieves competitive results on various machine learning tasks.
In this paper, we study the representation of neural networks from the view of kernels. We first define the Neural Fisher Kernel (NFK), which is the Fisher Kernel applied to neural networks. We show that NFK can be computed for both supervised and unsupervised learning models, which can serve as a unified tool for representation extraction. Furthermore, we show that practical NFKs exhibit low-rank structures. We then propose an efficient algorithm that computes a low rank approximation of NFK, which scales to large datasets and networks. We show that the low-rank approximation of NFKs derived from unsupervised generative models and supervised learning models gives rise to high-quality compact representations of data, achieving competitive results on a variety of machine learning tasks.