Logistic Regression Transfer Learning Mohammad Masum Ph D

Logistic Regression & Transfer Learning Mohammad Masum Ph. D. Student Institute of Analytics and Data Science Kennesaw State University

Logistic Regression • Mnist Digit Data • Binary classification; Label 7 & 8 • Shape 8, 000 x 784

Logistic Regression • Mnist Digit Data • Binary classification • Shape 8, 000 x 784

Transfer Learning • Transfer learning is a machine learning method where a model developed for a task is reused as the starting point for a model on a second task • Three Possible Benefits: • Higher start • Higher slope • Higher asymptote • The benefits of using transfer learning is not obvious

Transfer Learning: Pre-trained Model Available Models in Keras Framework • Models for image classification with weights trained on Image. Net • Xception • VGG 16 • Built by: Oxford Visual Geometry • VGG 19 Group • Res. Net 50 • Total 16 layers • Inception. V 3 • Given image find object name in • Inception. Resnet. V 2 the image • Mobile. Net • It can detect any one of 1000 images • Dence. Net • It take input image size 224 x 3 • Nas. Net (RGB image) • Mobile. Net. V 2

Pre-Trained Model Number of Features Feature Extraction

VGG 16 Architecture

VGG 16 Architecture

Last Conv Layer Flatten Data Feature Extraction Mnist Image Data Representation

Logistic regression with Extracted Features Data

Logistic regression with Extracted Features Data Possible reasons that VGG 16 does not perform well: • VGG 16 is trained for 3 -channel RGB images while mnist digit data is 1 -channel gray scale • Background of images • VGG 16 trained for 1, 000 objects label