Multimodality in meta-learning: A comprehensive survey

Meta-learning has gained wide popularity as a training framework that is more data-efficient than traditional machine learning methods. However, its generalization ability in complex task distributions, such multimodal tasks, not been thoroughly studied. Recently, some studies on multimodality-based meta-learning have emerged. This survey provides comprehensive overview of the landscape terms methodologies and applications. We first formalize definition multimodality, along with research challenges this growing field, how to enrich input few-shot (FSL) or zero-shot (ZSL) scenarios generalize models new tasks. then propose taxonomy discuss typical algorithms tasks systematically. investigate contributions related papers summarize them by our taxonomy. Finally, we potential directions for promising field.