What is xiaoyang-machine-learning?

xiaoyang-machine-learning is a Chinese-language educational website for learning machine learning and deep learning concepts. The entire site is contained in a single HTML file that opens directly in a browser without any installation, and is also hosted online for easy access. The site targets university students and AI enthusiasts at the beginning of their learning journey. It covers a progression from foundational concepts to modern techniques. The content starts with machine learning basics such as data, features, labels, models, loss functions, overfitting, and evaluation metrics, then moves to classical algorithms including linear regression, decision trees, random forests, gradient boosting variants (GBDT, XGBoost, LightGBM), SVM, and clustering methods. It continues into deep learning (neural networks, backpropagation, optimizers, batch normalization, and the PyTorch training workflow) and modern AI topics including Transformers, attention mechanisms, embeddings, retrieval-augmented generation, AI agents, LoRA fine-tuning, multimodal systems, and diffusion models. Beyond theory, the site includes sections on research skills: reading papers, designing experiments, running ablation studies, analyzing errors, and writing up results. There is also a section dedicated to preparing for job interviews at large technology companies, covering common algorithm interview questions, project presentation templates, and real business scenario analysis. The README suggests a three-pass reading approach: first for intuition and examples without stopping at formulas, second for formulas and code details, third for research insights and interview content. It encourages learners to ask themselves three questions about every algorithm they study: what it is, why it is still used, and what real problem it solves. The project is written in Chinese and intended for a Chinese-speaking academic audience. Contributions such as additional examples, paper notes, and interview questions are welcome.