Deep Learning for Vision - customised

There are no specific requirements needed to attend this course.

Deep Learning vs Other Methods

• When Deep Learning is suitable
• Limits of Deep Learning
• Machine Learning Limitations
• Comparing Performance and costs of different methods

Deep Learning Fundamentals

• Network Architectures (and Layers)
• Forward / Backward models
• Loss functions
• Layer types
• Convolutional Neural Networks (CNNs)
• Recurrent Neural Networks (RNNs) and LSTMs
• Preprocessing methods
• Postprocessing techniques

Design and Deployment

• Data preparation
• Selecting pre and post processing methods
• Designing products with deep learning
• Planning for deployment at scale
• Assessing the results of layers and CNNs
• Monitoring training quantitatively and graphically
• Optimization strategies

Models

• Regression - % Density of a Mammogram
• Classification - Severity Categorization
• Stretch: Primer on text analysis with deep learning