
Mingming Gong
Research Fellow Email: gongmingnjuatgmaildotcom 




Causal Inference
Modeling and Estimation of statistical dependences is a core problem of statistical machine learning. The development of machine learning methods has enabled people in various fields to perform accurate prediction based on the discovered complex nonlinear dependences from data. However, when one aims to predict the outcomes of interventions, one need to figure out whether the dependences between variables encode causal information. Causal inference is to develop reliable theories and algorithms for causal discovery from empirical data. In recent years, we have developed novel methods to discover causal relationships from time series. 



Domain Adaptation
Standard supervised learning relies on the assumption that both training and test data are drawn from the same distribution. However, this assumption is likely to be violated in practice if the training and test data are sampled under different conditions. Domain adaptation approaches aim to solve the domainmismatch problems by transferring knowledge between domains. To successfully transfer knowledge between domains, one need to capture the underlying causal mechanism, or the data generating process. In particular, for domain adaptation, one would be interested in what types of information are invariant, what types of information change, and how they change across domains. To this end, we have addressed the domain adaptation problem using causal models to characterize how the distribution changes between domains. 

