Scalable Kernel Learning Via the Discriminant Information

• Published in: ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 3152 - 3156
• Main Authors: Al, Mert, Hou, Zejiang, Kung, Sun-Yuan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.05.2020
• Subjects: classification; discriminant analysis; Kernel; kernel approximation; Kernel learning; Optimization; scalable learning; Speech processing; Supervised learning; Task analysis; Training
• DOI: 10.1109/ICASSP40776.2020.9053142

Kernel approximation methods create explicit, low-dimensional kernel feature maps to deal with the high computational and memory complexity of standard techniques. This work studies a supervised kernel learning methodology to optimize such mappings. We utilize the Discriminant Information criterion, a measure of class separability with a strong connection to Discriminant Analysis. By generalizing this measure to cover a wider range of kernel maps and learning settings, we develop scalable methods to learn kernel features with high discriminant power. Experimental results on several datasets showcase that our techniques can improve optimization and generalization performances over state of the art kernel learning methods.