Title: Knowledge Transfer Using Latent Variable Models

Speaker: Ayan Acharya, University of Texas, Austin

Abstract: In several applications, scarcity of labeled data is a challenging problem that hinders the predictive capabilities of machine learning algorithms. Additionally, the distribution of the data changes over time rendering models trained with older data less capable of discovering useful structure from the newly available data. Transfer learning is a convenient framework to overcome such problems where the learning of a model specific to a domain can benefit the learning of other models in other domains through either simultaneous training of domains or systematic sequential transfer of knowledge from one domain to the others.
In this talk, variety of methods for achieving transfer learning are explored where a hidden layer is shared across different domains in hierarchical latent variable based models. Applications of such frameworks in problems like text classification, object recognition from images, network modeling for community detection and evolution of count values vectors and matrices have shown excellent results so far. In particular, the strength of non-parametric Poisson factorization will also be explored in the context of knowledge transfer and modeling of large scale count data prevalent in text mining, image analysis and recommender systems.