Machine learning models need numerical representations of language, and many language tasks also require attention to word order. This course introduces word embeddings and recurrent neural networks as approaches to representing words and processing text sequences. You will examine the challenges of using text in machine learning, how word embeddings are learned, and how pretrained embeddings can support text classification.
You will then study recurrent neural networks, their use in classification, and the roles of pretraining and fine-tuning when adapting a model to a task. By the end of the course, you will be able to use word embeddings in classification, explain how recurrent networks process sequences, and apply pretrained neural models through fine-tuning.