Neural Generalization of Multiple Kernel Learning
This work addresses the performance gap between MKL and deep learning for researchers in kernel methods, though it is incremental as it builds on existing MKL frameworks.
The paper tackled the problem of Multiple Kernel Learning (MKL) having lower complexity and accuracy compared to deep learning models by proposing a neural generalization that extends MKL to multi-layer networks with nonlinear activations, resulting in improved complexity and higher recognition accuracy on benchmarks.
Multiple Kernel Learning is a conventional way to learn the kernel function in kernel-based methods. MKL algorithms enhance the performance of kernel methods. However, these methods have a lower complexity compared to deep learning models and are inferior to these models in terms of recognition accuracy. Deep learning models can learn complex functions by applying nonlinear transformations to data through several layers. In this paper, we show that a typical MKL algorithm can be interpreted as a one-layer neural network with linear activation functions. By this interpretation, we propose a Neural Generalization of Multiple Kernel Learning (NGMKL), which extends the conventional multiple kernel learning framework to a multi-layer neural network with nonlinear activation functions. Our experiments on several benchmarks show that the proposed method improves the complexity of MKL algorithms and leads to higher recognition accuracy.