CNNs Avoid Curse of Dimensionality by Learning on Patches

no code implementations • 22 May 2022 • Vamshi C. Madala, Shivkumar Chandrasekaran, Jason Bunk

A priori theories explaining the generalization performances of deep neural networks have mostly ignored the convolutionality aspect and do not specify why CNNs are able to seemingly overcome curse of dimensionality on computer vision tasks like image classification where the image dimensions are in thousands.

Data Augmentation Image Classification

Predicting Generalization in Deep Learning via Local Measures of Distortion

no code implementations • 13 Dec 2020 • Abhejit Rajagopal, Vamshi C. Madala, Shivkumar Chandrasekaran, Peder E. Z. Larson

We study generalization in deep learning by appealing to complexity measures originally developed in approximation and information theory.

Quantization