Trustworthy Deep Learning

This course on Trustworthy Deep Learning aims to teach learners about the challenges of robust deep learning, including algorithmic bias, uncertainty in deep learning, and methods for debiasing variational autoencoders. The course covers topics such as class and latent feature imbalance, aleatoric and epistemic uncertainty, and the application of risk-aware AI through Themis AI and Capsa. The teaching method includes lectures with a focus on theoretical concepts and real-world applications. This course is intended for individuals interested in deep learning, AI ethics, and ensuring the trustworthiness of AI systems.