Infinite-Dimensional Feature Interaction
This addresses a fundamental bottleneck in neural network design for improving information transformation across various ML/AI applications.
The paper tackles the limitation of neural networks capturing only finite-dimensional feature interactions by introducing InfiNet, which enables infinite-dimensional interactions using an RBF kernel, achieving new state-of-the-art performance with significant enhancements.
The past neural network design has largely focused on feature representation space dimension and its capacity scaling (e.g., width, depth), but overlooked the feature interaction space scaling. Recent advancements have shown shifted focus towards element-wise multiplication to facilitate higher-dimensional feature interaction space for better information transformation. Despite this progress, multiplications predominantly capture low-order interactions, thus remaining confined to a finite-dimensional interaction space. To transcend this limitation, classic kernel methods emerge as a promising solution to engage features in an infinite-dimensional space. We introduce InfiNet, a model architecture that enables feature interaction within an infinite-dimensional space created by RBF kernel. Our experiments reveal that InfiNet achieves new state-of-the-art, owing to its capability to leverage infinite-dimensional interactions, significantly enhancing model performance.