lizhuofan-curry/lihongyiml-2021-2022

This repository is a personal archive of course materials from Professor Hung-yi Lee's Machine Learning classes, covering the 2021 and 2022 Spring semesters at National Taiwan University. It is not the official course repository, it is one student's collection of the publicly available slides and homework, organized for easier personal reference. The README is explicit that anyone wanting the current schedule or official versions should check the original course homepages instead. The 2021 section holds lecture PDFs and slide decks across thirteen numbered lectures, covering topics like introduction and regression, deep learning basics, convolutional networks and self-attention, transformers, generative models such as GANs, self-supervised learning methods like BERT, explainable AI, adversarial attacks, domain adaptation, reinforcement learning, and meta learning. A CSV file lists every original file name and its source download link. The 2022 section mirrors an existing official homework repository, with lecture slides across fifteen lectures covering similar ground plus additional topics like anomaly detection and network compression, along with three homework assignments. Each homework folder includes the code, the data used, the results produced when running it, and notes on how to reproduce those results. The three included assignments cover a COVID-19 regression task, a phoneme classification task, and a food image classification task recognizing eleven food categories. Because some of the slide files are large, the repository uses Git LFS to store them, so anyone cloning it needs Git LFS installed beforehand, or needs to run a separate pull command afterward if they already cloned without it. The README is careful to note that the lecture materials themselves belong to the original professor, instructors, and course team, and that the repository's own MIT license only covers content the repository author actually created, not the third-party course slides.