Volha Petukhova

The independence of dimensions in multidimensional dialogue act annotation

This paper presents empirical evidence for the orthogonality of the DIT++ multidimensional dialogue act annotation scheme, showing that the ten dimensions of communication which underlie this scheme are addressed independently in natural dialogue.

A Hierarchical Unification of LIRICS and VerbNet Semantic Roles

This research compares several of the thematic roles of Verb Net (VN) to those of the Linguistic Infrastructure for Interoperable Resources and Systems (LIRICS). The purpose of this comparison is to develop a standard set of thematic roles that would be suited to a variety of natural language processing (NLP) applications. We draw from both resources to construct a unified set of semantic roles that will replace existing VN semantic roles. Through the process of comparison, we find that a hierarchical organization of coarse-grained, intermediate and fine-grained roles facilitates mapping between semantic resources of differing granularity and allows for flexibility in how VN can be used for diverse NLP applications, thus, we propose a hierarchical taxonomy of the unified role set. The comparison and subsequent development of the hierarchy reveals a level of granularity shared by both resources, which could be further developed into a standard set of thematic roles for the International Organization for Standardization (ISO).

Metalogue: A Multiperspective Multimodal Dialogue System with Metacognitive Abilities for Highly Adaptive and Flexible Dialogue Management

This poster paper presents a high-level description of the Metalogue project that is developing a multi-modal dialogue system that is able to implement interactive behaviors that seem natural to users and is flexible enough to exploit the full potential of multimodal interaction. We provide an outline of the initial work undertaken to define a an open architecture for the integrated Metalogue system. This system includes components that are necessary for the implementation of the processing stages for a variety of application domains: initialization, training, information gathering, orchestration, multimodality, dialogue management, speech recognition, speech synthesis and user modelling.

Towards a Multidimensional Semantics of Discourse Markers in Spoken Dialogue

The literature contains a wealth of theoretical and empirical analyses of discourse marker functions in human communication. Some of these studies address the phenomenon that discourse markers are often multifunctional in a given context, but do not study this in systematic and formal ways. In this paper we show that the use of multiple dimensions in distinguishing and annotating semantic units supports a more accurate analysis of the meaning of discourse markers. We present an empirically-based analysis of the semantic functions of discourse markers in dialogue. We demonstrate that the multiple functions, which a discourse marker may have, are automatically recognizable from utterance surface-features using machine-learning techniques.

Dialogue Act Annotation with the ISO 24617-2 Standard

This chapter describes recent and ongoing annotation efforts using the ISO 24617-2 standard for dialogue act annotation. Experimental studies are reported on the annotation by human annotators and by annotation machines of some of the specific features of the ISO annotation scheme, such as its multidimensional annotation of communicative functions, the recognition of each of its nine dimensions, and the recognition of dialogue act qualifiers for certainty, conditionality, and sentiment. The construction of corpora of dialogues, annotated according to ISO 24617-2, is discussed, including the recent DBOX and DialogBank corpora.

Incremental Recognition and Prediction of Dialogue Acts

This chapter is concerned with the incremental understanding of utterances in spoken dialogue, with a focus on how their intended (possibly multiple) communicative functions can be recognized in a data-oriented way on the basis of observable features of communicative behaviour. An incremental, token-based approach is described which combines the use of local classifiers, that exploit local utterance features, and global classifiers that use the outputs of local classifiers applied to previous and subsequent tokens. This approach is shown to result in excellent dialogue act recognition scores for unsegmented spoken dialogue. This can be seen as a significant step forward towards the development of fully incremental, on-line methods for computing the meaning of utterances in spoken dialogue.

Observing, coaching and reflecting: Metalogue - A multi-modal tutoring system with metacognitive abilities

The Metalogue project aims to develop a multi-modal, multi-party, dialogue system with metacognitive abilities that will advance our understanding of natural conversational human-machine interaction, and interfaces that incorporate multimodality into virtual and augmented reality environments. In this paper we describe the envisaged technical system, the learning contexts it is being developed to support and the pedagogical framework in which it is proposed user interactions will take place. This includes details of the system-generated learner feedback provided both in-performance and post-performance.We then move on to explain what has been achieved so far in terms of the integrated system pilots, and finally we discuss three key challenges the Metalogue researchers are currently working to overcome.

Modelling Argumentative Behaviour in Parliamentary Debates: Data Collection, Analysis and Test Case

In this paper we apply the information state update (ISU) machinery to tracking and understanding the argumentative behaviour of participants in a parliamentary debate in order to predict its outcome. We propose to use the ISU approach to model the arguments of the debaters and the support/attack links between them as part of the formal representations of a participant’s information state. We first consider the identification of claims and evidence relations to their premises as an argument mining task. It is not sufficient, however, to indicate what relations occur without establishing how these relations are created and verified during the interaction. For this purpose the model requires a detailed specification of the creation, maintenance and use of shared beliefs. The ISU model provides procedures for incorporating beliefs and expectations shared between speaker and hearers in the tracking model. To evaluate the content of the tracked information states, we compare them to those of the human ‘concluder’ who wraps up a debate, stating the claims which the majority of the debaters have agreed on.

Understanding Questions and Extracting Answers: Interactive Quiz Game Application Design

The paper discusses two key tasks performed by a Question Answering Dialogue System (QADS): user question interpretation and answer extraction. The system represents an interactive quiz game application. The information that forms the content of the game is concerned with biographical facts of famous people’s life. The process of a question classification and answer extraction is performed based on a domain-specific taxonomy of semantic roles and relations computing the Expected Answer Type (EAT). Question interpretation is achieved performing a sequence of classification, information extraction, query formalization and query expansion tasks. The expanded query facilitates the search and retrieval of the information. The facts are extracted from Wikipedia pages by means of the same set of semantic relations, whose fillers are identified by trained sequence classifiers and pattern matching tools, and edited to be returned to the player as full-fledged system answers. The results (precision of 85% for the EAT classification of both in questions and answers) show that the presented approach fits the data well and can be considered as a promising method for other QA domains, in particular when dealing with unstructured information.