Creativity Inspired Zero-Shot Learning
This work addresses the problem of improving discriminative power in zero-shot learning for AI systems by drawing from psychology, offering a novel approach that could enhance performance in tasks with unseen categories, though it appears incremental as it builds on existing ZSL methods.
The paper tackles zero-shot learning (ZSL) by modeling the visual learning process of unseen categories with inspiration from human creativity, introducing a learning signal that explores unseen space with hallucinated class-descriptions and encourages deviation from seen classes while allowing knowledge transfer. Empirically, it shows consistent improvement over state-of-the-art by several percent on benchmarks like CUB and NABirds for generalized ZSL from noisy text, and also demonstrates advantages on attribute-based ZSL on datasets such as AwA2, aPY, and SUN.
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.