Neural Nets for NLP - Models with Latent Random Variables

This course covers the following learning outcomes and goals: - Understanding the differences between generative vs. discriminative models and deterministic vs. random variables. - Exploring Variational Autoencoders and their applications in Natural Language Processing (NLP). - Learning how to handle discrete latent variables in neural networks. - Gaining insights into examples of Variational Autoencoders in NLP. The course teaches the following individual skills or tools: - Deep Structured Latent Variable Models - Variational Inference techniques - Re-parameterization Trick for handling sampling issues - Training difficulties and solutions in Variational Autoencoders The teaching method of the course includes lectures, quizzes, practical exercises, and examples to illustrate concepts. The intended audience for this course is individuals interested in Neural Networks for NLP, particularly those looking to delve into advanced topics such as Variational Autoencoders and handling latent variables in neural network models.