Machine Learning Area

The Machine Learning Area at Microsoft Research Asia pushes the frontier of machine learning from the perspectives of theory, algorithms, and applications. Our research interests cover deep learning, reinforcement learning, graph learning, Boosting trees, online learning, pretraining, dynamics learning, and learning theory. In addition, we are also making active explorations on AI for Science (including biology, physics, sustainability) and AI for Industry (including finance, supply chain, and healthcare), with the mission to empower scientists and industry practitioners with our machine learning technologies (see our overall Research for more details). We have published many highly cited papers on top conferences and journals, transferred many technologies to Microsoft products and services, and helped many external partners achieve successful digital transformations. We have also released several open-sourced toolkits, such as LightGBM, LigthLDA, Microsoft Graph Engine, MARO, Qlib, and FOST, which attracted a lot of attention from the open-source community, and received over 30K stars on Github in total.

微软亚洲研究院机器学习领域从理论、算法、应用等不同层面推动机器学习的前沿。我们的研究兴趣包含：深度学习、强化学习、图学习、梯度提升树、在线学习、预训练、动态学习、学习理论等。同时，我们也在积极探索人工智能在自然科学和产业应用中的价值，从而为科学工作者和传统工业赋能（具体见研究概况）。在过去的十几年间，我们在顶级国际会议和期刊上发表了大量被高度引用的高质量论文，向微软的产品部门转化了大量核心技术，并帮助众多的企业合作伙伴实现了数字化转型。我们也向开源社区贡献了大量高质量开源工具，例如 LightGBM、LigthLDA、微软图引擎，多智能体资源优化平台“群策 MARO“，业内首个AI量化投资平台“微矿Qlib”，以及最新的时空预测平台»FOST»。这些工具受到开源社区的广泛关注，已在Github上累计收获三万余颗星。