MLJS08 - Convolutional Neural Network (CNN) modelling with Azure Machine Learning Service

Outline

• Construct and implement deep CNN model (incorporating convolutional and pooling layers with dropout) using Plain Tensorflow with Google’s open-source Dogs and Cats images
• Transfer learning from established CNN architectures (LeNet-5, AlexNet, GoogLeNet, ResNet) with Google’s open-source Dogs and Cats images
• Fine-tune hyperparameters (batch-size, learning rate, number of epochs, number of hidden layers, number of neurons per hidden layer, activation functions) of CNN model in Tensorflow with Google’s open-source Dogs and Cats images
• Construct and implement deep CNN model (incorporating convolutional and pooling layers with dropout) using Plain Tensorflow with Shopee’s open-source mobile and fashion product images

Requirements

• Basic knowledge of Python Programming
• Participants who are joining this series for the first-time should refer to this document
• Participants will need to bring their own laptops

Mode of Training

Classroom

Workshop Format

Lecture - 1 hr | Hands-on tutorial - 1.25 hr | Q & A – 0.5 hr

Trainer

Alvin Chew is currently a Microsoft Cloud Research Software Fellow and a final-year PhD candidate in School of Civil and Environmental Engineering, Nanyang Technological University.