This course provides an in-depth exploration of how natural language understanding and generation is used to improve learning efficiency and generalization in machine learning, focusing on seminal publications both theoretical and applied. We will cover topics that include meaning representation, knowledge-base population, question-answering, in-context learning, and planning. While there are no formal prerequisites, a background in machine learning and natural language processing is highly recommended. Upon completion, students will have a solid understanding of machine learning and natural language processing crucial for research in NLP for ML.
All late submissions are handled through official Short Term Absences and Verification of Illness. Late days are not granted outside of official STA/VIF channels.
This course is not a "bird class" where you can easily get by with minimal effort or rely on tools like GPT to do the work for you (I am very good at prompting and evaluating LLMs). Our focus is on engaging deeply with scientific research, which requires critical reading, thoughtful analysis, and active participation. If you are not willing to invest the time and mental energy to carefully read and think critically about academic papers, this course is not the right fit for you - you will not have a good time and you will worsen the learning experience for your fellow classmates! We are here to learn the science thoroughly and develop genuine understanding.