no code implementations • 17 Apr 2023 • Tomas McKelvey
A regression model with more parameters than data points in the training data is overparametrized and has the capability to interpolate the training data.
no code implementations • 21 Oct 2019 • Yinan Yu, Samuel Scheidegger, Tomas McKelvey
During the analysis, we observe that the effective rank of the vectorized Conv2D filters decreases with respect to the increasing depth in the network, which then leads to the implementation of the Depthwise Convolutional Eigen-Filter (DeCEF) layer.
no code implementations • 21 Oct 2019 • Yinan Yu, Tomas McKelvey
An extension to a hierarchical learning structure is also proposed to further improve the classification performance, where on each layer, the CLASMK is computed based on a selected "marginal" subset and feature vectors are constructed by concatenating the features from all previous layers.