Overview

Given the increased interest and requirements to explain decisions made by agents there is potential for approaches and techniques from various domains to provide novel approaches to supporting dialogue and explainability within the HAI context.

The International Workshop on Dialogue, Explanation and Argumentation in Human-Agent Interaction (DEXAHAI) will focus on the approaches, concepts and applications relevant to supporting dialogue and explainability in the Human-Agent Interaction (HAI) context. We focus on the following questions:

• How to develop methods to facilitate dialogue between humans and agents?
• How to derive arguments given such dialogues?
• How to explain the decisions made by the intelligent systems and how humans interact with such systems?

Workshop Goals

The aim of the International Workshop on Dialogue, Explanation and Argumentation in Human-Agent Interaction (DEXAHAI) is to provide a forum to discuss and advance the potential synergy and collaborations within these domains. Specific topics for this workshop include but are not limited to:

• Collaborative Multi-agent systems based on argumentation
• Dialogue between agents
• Explainable AI
• Agreement technologies (negotiation and argumentation)
• Studies in Explainability in Human-Agent Interaction
• Trust modeling for Multi-agent decision making;
• Feedback on presenting argumentation reasoning to support human agent interaction
• Empirical evaluations of argumentation systems involving collaborations between humans and agent

This workshop will cover the use of computational argumentation approaches to answer these questions. The workshop will also include an invited talk.

Workshop Schedule and Links to Papers

13:30 - 13:40 - Opening

13:40 - 14:05 - Oana Cocarascu, Francesca Toni and Catalina Popescu - A platform for crowdsourcing corpora for argumentative dialogue

14:05 - 14:30 - Nir Oren, Quratul-Ain Mahesar and Wamberto Vasconcelos - Detecting Preferences for Dialogue

14:30 - 14:55 - Nathan Duran and Steve Battle - Conversation Analysis Structured Dialogue for Multi-Domain Dialogue Management

14:55 - 15:20 - Isabel Sassoon, Nadin Kokciyan, Simon Parsons and Elizabeth Sklar - Towards the use of Commitments in Multi-agent decision support systems

15:30 - 16:00 - Coffee/Tea Break

16:00 - 17:00 - Invited talk: Timothy Norman, University of Southampton - Understanding the benefits of argumentation in support of human reasoning: A study with expert intelligence analysts

17:00 - 17:25 - Khaled Belahcene, Christophe Labreuche, Nicolas Maudet, Vincent Mousseau and Wassila Ouerdane - Challenges in Interactive Explanation and Recommendation for Decision Support

17:25 - 17:50 - Oana Cocarascu, Kristijonas Cyras, Antonio Rago and Francesca Toni - Explaining with Argumentation Frameworks Mined from Data

17:50 - 18:00 - Closing

The workshop will take place in Room 1023 in the Murray Building (Building 58) at Highfield campus, the main campus of the University of Southampton. A map of the campus is provided here.

Submission Format

Submissions are in the format of short papers (2 to 4 pages). These should be submitted using the main ACM HAI conference template link to template .

Submissions will be via EasyChair - submission link

Important Dates

• Submission deadline: 5th November 2018 (11:59PM TAHT - Tahiti Time) EXTENDED!
• Notification of acceptance: 15th of November 2018 EXTENDED!
• Workshop date: 15th December 2018 ( Location and time TBC)

Organising Committee

• Isabel Sassoon (King's College London)
• Nadin Kokciyan (King's College London)
• Simon Parsons (King's College London)
• Timothy Norman (University of Southampton)