A Survey on Cross-Modal Embedding

This document provides a survey of research on cross-modal embedding and retrieval between different media types such as text, images, video, and audio. It discusses several key areas including cross-modal retrieval which aims to retrieve relevant items across media types, audio-visual embedding to learn joint representations of audio and video, and applications such as sound localization, generat

[knok]

異なるモーダル間での関連性を持つEmbedding獲得。adversarialな手法やconsistencyが有効