# Bayesian Zero-Shot Learning

## TL;DR

This paper introduces a Bayesian hierarchical model for zero-shot learning that leverages meta-classes and priors to improve recognition of unseen classes, demonstrating superior performance on benchmark datasets including ImageNet.

## Contribution

It presents a novel Bayesian hierarchy with meta-classes for generalized ZSL, effectively blending data likelihood with priors to enhance unseen class recognition.

## Key findings

- Outperforms existing methods on seven benchmark datasets.
- Achieves state-of-the-art results on large-scale ImageNet.
- Provides a flexible hyperparameter tuning mechanism for trade-offs.

## Abstract

Object classes that surround us have a natural tendency to emerge at varying levels of abstraction. We propose a Bayesian approach to zero-shot learning (ZSL) that introduces the notion of meta-classes and implements a Bayesian hierarchy around these classes to effectively blend data likelihood with local and global priors. Local priors driven by data from seen classes, i.e. classes that are available at training time, become instrumental in recovering unseen classes, i.e. classes that are missing at training time, in a generalized ZSL setting. Hyperparameters of the Bayesian model offer a convenient way to optimize the trade-off between seen and unseen class accuracy in addition to guiding other aspects of model fitting. We conduct experiments on seven benchmark datasets including the large scale ImageNet and show that our model improves the current state of the art in the challenging generalized ZSL setting.