A study on annotation of utterance units and their functions in dialogs

(1)提出了一种用于日语对话中话语单元的标注方案，通过该方案可以对具有两种不同粒度的话语，长和短的话语单位，确定，已提出并验证其功效。(2)严格和一致的程序的话语单位注释已经制定和记录。(3)对于一组对话，共40分钟的录音，长和短的话语单位已被注释，连同其他语言和intermittent信息，如形态，韵律，和话轮转换信息。(4)注释的响应令牌，这是不包括的话语单元注释方案，也被记录和进行相同的数据集。(5)基于这些注释的数据，已经获得了一些新的和先进的发现。

(1) An annotation scheme for utterance-units in Japanese dialogs, by which utterances with two different granularities, i.e., Long and Short Utterance-Units, are identified, has been proposed and verified its efficacy.(2) Strict and consistent procedures for the utterance-unit annotation has been developed and documented.(3) For a set of dialogs consisting of a total of 40-minute recordings, long and short utterance-units have been annotated, together with other linguistic and interactional information such as morphological, prosodic, and turn-taking information.(4) Annotation of response tokens, which is not covered by the utterance-unit annotation scheme, has also been documented and conducted for the same set of data.(5) Based on these annotated data, a number of novel and advanced findings have been obtained.