李洋博士在营销数据模型、人工智能算法等方面的研究成果已发表在Management Science、Marketing Science、Journal of Marketing Research等管理类国际A级学术期刊上，常在美国和欧洲的学术机构做关于数据模型的演讲。在长江商学院李洋讲授EMBA、EE、FMBA、MBA等项目课程，曾为腾讯、百度、永辉超市、海尔等企业提供营销咨询，并持有医学图像处理的美国专利。
In this paper we develop anovel covariate-guided heterogeneous supervised topic model that uses productcovariates, user ratings and product tags to succinctly characterize productsin terms of latent topics, and speciﬁes consumer preferences via these topics.Recommendation contexts also generate big data problems stemming from datavolume, variety and veracity, as in our setting that includes massive textualand numerical data. We therefore develop a novel stochastic variationalBayesian (SVB) framework to achieve fast, scalable and accurate estimation insuch big data settings, and apply it to a MovieLens dataset of movie ratingsand semantic tags. We show that our model yields interesting insights aboutmovie preferences and predicts much better than a benchmark model that onlyuses product covariates. We showcase how our model can be used for targetingrecommendations to particular users and illustrate its use in generatingpersonalized search rankings of relevant products.