Hierarchical Argumentation Structure for Persuasive Argumentative Dialogue Generation

Abstract

Argumentation is a process of reaching a consensus through premises and rebuttals. If an artificial dialogue system can perform argumentation, it can improve users' decisions and ability to negotiate with the others. Previously, researchers have studied argumentative dialogue systems through a structured database regarding argumentation structure and evaluated the logical consistency of the dialogue. However, these systems could not change its response based on the user's agreement or disagreement to its last utterance. Furthermore, the persuasiveness of the generated dialogue has not been evaluated. In this study, a method is proposed to generate persuasive arguments through a hierarchical argumentation structure that considers human agreement and disagreement. Persuasiveness is evaluated through a crowd sourcing platform wherein participants' written impressions of shown dialogue texts are scored via a third person Likert scale evaluation. The proposed method was compared to the baseline method wherein argument response texts were generated without consideration of the user's agreement or disagreement. Experiment results suggest that the proposed method can generate a more persuasive dialogue than the baseline method. Further analysis implied that perceived persuasiveness was induced by evaluations of the behavior of the dialogue system, which was inherent in the hierarchical argumentation structure.

A study of interactive robot architecture through the practical implementation of conversational android

“…Task-oriented dialogues such as voice guidance ( Adachi et al, 2003 ) and discussion ( Sakai et al, 2020 ) have clear goals, and the dialogue strategy is designed to achieve the goal. In these dialogues, the conveyed information through the dialogue is important, and the existing studies focus on the design of the dialogue strategy to obtain and provide information necessary to achieve the goal.…”

Integrating Mathematics and Science Teaching in the Context of Education for Sustainable Development: Design and Pilot Implementation of a Teaching-Learning Sequence about Air Quality with Pre-Service Primary Teachers

“…The fact that during the implementation students statistically analyzed the evolution of the leading air pollutants in the city of Vitoria-Gasteiz (O 3 , PM 2.5 and PM 10 ) and learned about descriptive statistical parameters and statistical data graphics helped them to interpret the evidence supporting the good air quality of the campus at UPV/EHU. The use of arguments is integral in decision-making processes [61], and will help pre-service teachers to actively participate in a democratic society [62]. Therefore, thanks to such learning experiences, pre-service teachers' scientific literacy improves, which will, in turn, improve their contributionsboth personal and professional-to the achievement of SDGs, particularly through the promotion of ESD.…”

Proceedings of the 22nd Annual Meeting of the Special Interest Group on Discourse and Dialogue

“…In addition to these perspectives, we need a filter for personas of dialog systems. Persona-aware dialog systems are important in that having a consistent persona makes a dialog system believable (Higashinaka et al, 2018) Figure 1 shows how we would like to automatically evaluate the appropriateness of a large number of arbitrary utterances and select utterances to be registered in the utterance database of a retrieval-based dialog system. Doing this is preferable for commercial use in terms of preventing unexpected utterances from being output.…”

“…Existing systems address a multitude of tasks, including full scale debates against a human debater (Slonim et al, 2021), persuasion (Chalaguine and Hunter, 2020) and customer support (Galitsky, 2019). Due to the complex nature of this type of interaction, many systems rely on topic-specific argument structures that encode arguments and their relations to ensure a consistent and challenging interaction (Rach et al, 2018b;Sakai et al, 2020). Despite the advantages on the formal side, this dependency limits the range of topics that can be discussed by a system as the required structures are often either annotated by hand (Rach et al, 2019;Sakai et al, 2018b) or acquired in time-consuming data collections .…”

“…Due to the complex nature of this type of interaction, many systems rely on topic-specific argument structures that encode arguments and their relations to ensure a consistent and challenging interaction (Rach et al, 2018b;Sakai et al, 2020). Despite the advantages on the formal side, this dependency limits the range of topics that can be discussed by a system as the required structures are often either annotated by hand (Rach et al, 2019;Sakai et al, 2018b) or acquired in time-consuming data collections . This limitation renders the corresponding systems inflexible, especially in comparison to recent data-driven approaches in domains like question answering .…”