Note: This code is based on the new split proposed by Y Xian et. al.

A Simple Exponential Family Framework for Zero-Shot Learning

Abstract:

Abstract. We present a simple generative framework for learning to predict previously unseen classes, based on estimating class-attribute-gated class-conditional distributions. We model each class-conditional distribution as an exponential family distribution and the parameters of the distribution of each seen/unseen class are defined as functions of the respective observed class attributes. These functions can be learned using only the seen class data and can be used to predict the parameters of the class-conditional distribution of each unseen class. Unlike most existing methods for zero-shot learning that represent classes as fixed embeddings in some vector space, our generative model naturally represents each class as a probability distribution. It is simple to implement and also allows leveraging additional unlabeled data from unseen classes to improve the estimates of their class-conditional distributions using transductive/semi-supervised learning. Moreover, it extends seamlessly to few-shot learning by easily updating these distributions when provided with a small number of additional labelled examples from unseen classes. Through a comprehensive set of experiments on several benchmark data sets, we demonstrate the efficacy of our framework.

Prerequisites

Matlab
vlfeat toolbox

Usage

cub.m
sun.m
awa1.m
awa2.m

Dataset

• AWA2: Animals with Attributes 2
• CUB: Caltech-UCSD Birds-200-2011
• SUN: SUN Attribute

Complete Datasets can be downloaded here. For more detail about train/test split please refer to our paper

Result

Here GFZSL-Trans represents the result in the transductive.

References

If you are using this work please refer the ECML-17 paper:

@inproceedings{verma2017simple,
  title={A simple exponential family framework for zero-shot learning},
  author={Verma, Vinay Kumar and Rai, Piyush},
  booktitle={Joint European Conference on Machine Learning and Knowledge Discovery in Databases},
  pages={792--808},
  year={2017},
  organization={Springer}
}

License

Code are released for non-commercial and research purposes only. For commercial purposes, please contact the authors.