Zero-Shot Learning via Latent Probabilistic Modeling


Abstract: In this talk we consider a version of the zero-shot recognition (ZSR) problem where seen class source and target domain data are provided. The goal during test-time is to accurately predict the class label of an unseen target domain instance based on revealed source domain side information (e.g. attributes) for unseen classes. We formulate ZSR as a binary prediction problem. Our resulting classifier is class-agnostic. It takes an arbitrary pair of source and target domain instances as input and predicts whether or not they come from the same class, i.e. whether there is a match. We model the posterior probability of a match since it is a sufficient statistic and propose a latent probabilistic model in this context. We accordingly develop two ways of parameterization of our probabilistic model: (1) Our first parameterization is based on viewing each source or target data instance as a mixture of seen class proportions and we postulate that the mixture patterns have to be similar if the two instances belong to the same unseen class. This perspective leads us to learning source/target embedding functions that map an arbitrary source/target domain data instance into a same semantic space where similarity can be readily measured. (2) Our second parameterization is a joint discriminative learning framework based on dictionary learning to jointly learn the parameters of our model for both domains. Note that many of the existing embedding methods can be viewed as special cases of our probabilistic model. Empirically our method is tested on several benchmark datasets and achieves the state-of-the-art on all the datasets.

Guest Lectuer: Dr. Zheng Ziming, Boston University, USA

Date and Time: 10:30am-11:45am, Jan. 8, 2016  

Location: Room 108, Administration Building, Yuquan Campus

Audience: Faculty/Staff, Students

Category: Lecture

Sponsor: College of Information Science & Electronic Engineering, Zhejiang University


Admission: Free