The DialogBank

The DialogBank Harry Bunt*, Volha Petukhova**, Andrei Malchanau**, Alex Fang*** and Kars Wijnhoven* *Tilburg Center for Cognition and Communication, Tilburg University, The Netherlands **Spoken Language Systems, Saarland University, Germany ***Dialogue Systems Group, City University of Hong Kong, Hong Kong SAR harry.bunt,kars.wijnhoven@uvt.nl; v.petukhova, a.malchanau@lsv.uni-saarland.de, alex.fang@cityu.edu.hk Abstract This paper presents the DialogBank, a new language resource consisting of dialogues with gold standard annotations according to the ISO 24617-2 standard. Some of these dialogues have been taken from existing corpora and have been re-annotated according to the ISO standard; others have been annotated directly according to the standard. The ISO 24617-2 annotations have been designed according to the ISO principles for semantic annotation, as formulated in ISO 24617-6. The DialogBank makes use of three alternative representation formats, which are shown to be interoperable. Keywords: dialogue annotation, ISO standard, interoper- ability 1. Introduction This DialogBank is a new language resource, developed at Tilburg University, which contains dialogues of various kind with gold standard dialogue act annotations accord- ing to the ISO 24617-2 standard. This standard builds on previously designed annotation schemes such as DAMSL, DIT++, MRDA, HCRC Map Task, Verbmobil, SWBD- DAMSL, and DIT.1 Most of these schemes have been used to construct annotated corpora, such as the Switchboard, HCRC Map Task, ICSI-MRDA, and DIAMOND corpora. For nearly all of these annotation schemes, dialogue act annotation consists of segmenting a dialogue into certain grammatical units and marking up each unit with one or more communicative function labels. ISO 24617-2 (like DIT++) supports semantically more complete annotation by additionally annotating the following aspects: 1. ’Dimension’, or category of semantic content: the annotation scheme supports multidimensional annota- tion, i.e. multiple communicative functions may be as- signed to dialogue segments; different from DAMSL and other multidimensional schemes, an explicitly de- fined notion of ‘dimension’ is used that corresponds to a certain category of semantic content. The ISO scheme distinguishes nine dimensions on empirical and theoretical grounds. ’Multidimensional segmentation’ is used: a dialogue is segmented in multiple ways, with functional seg- ments in each relevant dimension. ‘Functional seg- ments’, defined as minimal stretches of behaviour that have a communicative function, are the units that di- alogue act annotations attach to. Functional segments are mostly shorter than turns, may be discontinuous, may overlap, and may contain parts contributed by dif- ferent speakers. A segment carrying a feedback func- tion may for instance overlap with a segment that car- ries a task-related function. 1See Allen & Core (1997); Bunt (2007); Shriberg et al. (2004); Anderson et al. (1991); Alexandersson et al. (1998); Jurafsky et al. (1997); and Bunt (1994; 2000), respectively. 2. ‘Qualifiers’ may be added for expressing that a dia- logue act is performed conditionally, with uncertainty, or with a particular sentiment. 3. Dependence relations are defined for expressing se- mantic relations between dialogue acts, e.g. for indi- cating which question is answered by a certain answer act, or which utterance a feedback act. responds to. 4. Rhetorical relations may be annotated to indicate e.g. that one dialogue act …

Content truncated to 500 words