A fundamental understanding of human perception and impressions evoked from human language can help with multimodal processing from robotics to multimedia content analysis. In this talk, I discuss our research on incorporating ideas from psycholinguistics into computing. First, I introduce our research on estimating the perceived imageability of a word based on analyzing image-text data. Second, I introduce our research on visualizing the mental images of spoken language.
Invited talk at RIKEN Guardian Robot Project (RIKEN GRP 研究会)