ISO-Standard Domain-Independent Dialogue Act Tagging for Conversational\n Agents

Abstract

Dialogue Act (DA) tagging is crucial for spoken language understanding\nsystems, as it provides a general representation of speakers' intents, not\nbound to a particular dialogue system. Unfortunately, publicly available data\nsets with DA annotation are all based on different annotation schemes and thus\nincompatible with each other. Moreover, their schemes often do not cover all\naspects necessary for open-domain human-machine interaction. In this paper, we\npropose a methodology to map several publicly available corpora to a subset of\nthe ISO standard, in order to create a large task-independent training corpus\nfor DA classification. We show the feasibility of using this corpus to train a\ndomain-independent DA tagger testing it on out-of-domain conversational data,\nand argue the importance of training on multiple corpora to achieve robustness\nacross different DA categories.\n