Construction of a Topic Database for Remote Conversation Support

Introduction

This paper presents a topic database for supporting conversation with people with such cognitive handicaps as aphasia. There are several conversation support systems proposed for people with aphasia to be used with Internet. For example, a supportive Web browser was developed for people with cognitive and communication impairments including aphasia by the European Union (EU) World Wide Augmentative and Alternative Communication (WWAAC) project (Poulson & Nicolle, 2004). Spaniol et al. (2006) developed a web learning environment called SOCRATES to support learning activities of a community of people with aphasia, therapists and others. To help the conversation of such people, Yasuda et al. (2007) created a word list called Easy Natural Conversation that is written in the HTML format and is available on the Internet. In this word list, the words are arranged according to hierarchical categories. By following the hierarchies, an appropriate word can be selected and shown. Based on this word list concept, a remote conversation support system was constructed (Aye et al., 2010). This system uses an instant messaging system to allow the focus of attention to be shared between remote places. In addition, an agent-based approach was proposed to make a support more flexible (Kuwabara, Shimode, & Miyamoto, 2010). To implement such a system, the methodology to develop an intelligent software agent is needed.

In the field of Cognitive Informatics (CI), the internal information processing mechanisms and processes of a human brain are investigated (Wang et al., 2011). Their underlying theories are abstract intelligent theories and denotational mathematics, and their engineering applications are considered in the field of Cognitive Computing (CC). These methodologies and models can lay the foundation for implementing an intelligent conversation support system to adapt to each user’s specific needs.

In related to the Cognitive Informatics and Cognitive Computing, a symbiotic computing was proposed whose main concept is the symbiosis of human and agent (computer). The symbiotic computing aims to bridge an e-Gap between Real Space (RS) and Digital Space (DS) by bringing social heuristics and cognitive functions into DS (Sugawara, Fujita, & Hara, 2007). Socialware (Hattori et al., 1999) is one of the core technologies of symbiotic computing, and is intended to form a basis of a multi-agent system to support various activities on network communities. The concept of socialware was applied to develop a support system for people with cognitive handicaps (Hattori et al., 2010). In the agent-based support system, typically a user agent is placed in corresponding to a human user. The user agents communicate with each other to provide a flexible support in an intelligent way. In such a system, each user agent needs to be designed carefully. The cognitive informatics and cognitive computing will provide a basis for implementing such an agent. In particular, a cognitive informatics reference model of autonomous agent systems (AAS) was proposed (Wang, 2009).