This repository contains official implementations of research in Bayesian Optimization, Reinforcement Learning, and Generative Models developed by Huawei Noah’s Ark Lab. It includes multiple projects such as HEBO for heteroscedastic evolutionary Bayesian optimization, various RL algorithms (SIMMER, PMDB), and LLM-related work like SparsePO for preference alignment and EM-LLM for episodic memory in LLMs. The repository serves as a toolkit for ML research and experimentation.