# Creativity Inspired Zero-Shot Learning

## TL;DR

This paper introduces a novel zero-shot learning method inspired by human creativity, using hallucinated class descriptions to improve recognition of unseen categories, achieving significant performance gains on multiple benchmarks.

## Contribution

We propose a creativity-inspired approach for zero-shot learning that explores unseen class space with hallucinated descriptions, enhancing discriminative power and transfer from seen classes.

## Key findings

- Consistent improvement over state-of-the-art on ZSL benchmarks.
- Effective on noisy text and attribute-based ZSL datasets.
- Advantage demonstrated across multiple datasets.

## Abstract

Zero-shot learning (ZSL) aims at understanding unseen categories with no training examples from class-level descriptions. To improve the discriminative power of zero-shot learning, we model the visual learning process of unseen categories with inspiration from the psychology of human creativity for producing novel art. We relate ZSL to human creativity by observing that zero-shot learning is about recognizing the unseen and creativity is about creating a likable unseen. We introduce a learning signal inspired by creativity literature that explores the unseen space with hallucinated class-descriptions and encourages careful deviation of their visual feature generations from seen classes while allowing knowledge transfer from seen to unseen classes. Empirically, we show consistent improvement over the state of the art of several percents on the largest available benchmarks on the challenging task or generalized ZSL from a noisy text that we focus on, using the CUB and NABirds datasets. We also show the advantage of our approach on Attribute-based ZSL on three additional datasets (AwA2, aPY, and SUN). Code is available.