CSC 5611 - Deep Learning

• Understand the process of backpropagation and its role in training deep networks 
• Implement a basic deep learning network library 
• Apply deep learning to a unique dataset and analyze performance 
• Compare options for training deep neural networks, for example, optimization methods, generalization methods, normalization, initialization methods  
• Apply transfer learning to leverage pre-trained models 
• Compare various modern network architectures, for example, convolutional neural networks (CNNs), encoder-decoders, generative adversarial networks (GANs), or transformers. Emphasis on topics can vary depending on the instructor’s specialties 
• Evaluate training methods and alternative architectures based on model accuracy, constraints, and training performance 
• Discuss the ethical implications of deep neural network applications 
• Compare the performance of the from-scratch deep neural network library implementation to an existing library when applied to a unique dataset 

• Machine learning
• Python fundamentals