Language Resources of the Future (a speech-based position paper)

Abstract

Current speech technology is well capable of processing the mappings between speech and text, but the challenge at the current time is to model the information flow in interactive speech, where propositional content is conveyed along with discourse controls to simultaneously signal the speaker?s cognitive attentional, intentional, and emotional states. Current applications of speech technology include general information services, customercare, robotics, games, and interactive media content, often making use of graphical interfaces which include an avatar, embodied conversational agent, or talking heads. Accordingly, the corpora that need to be collected for future speech technology research must be multi-modal, multicultural, and multilingual, incorporating material for the research of interpersonal communication strategies, and speech-related bodily movements, as well as the characteristics of interactive speech itself.