Back to search

Aristotle University Of Thessaloniki (AUTH)

Web-based conversational agents for collaborative learning support

Abstract

dc:description

Over the past decades, great strides have been made towards the development of software agents that engage in natural language interactions with the learner. Intelligent conversational agents have been developed and used repeatedly as pedagogical tools for fostering students’ motivation and engagement in individual learning settings. Research has also provided evidence that conversational agents can effectively be used to support peer discourse in the context of computer-supported collaborative learning (CSCL) activities. Through the lens of CSCL, this work explores a new generation of adaptable conversational agents employing domain-independent intervention strategies, which originate from the classroom discourse framework of academically productive talk (APT). Based on teachers’ exploration of core features of constructive classroom conversations, APT highlights a set of useful discussion facilitation strategies and norms that can enhance learning and promote students’ equitable and reasoned participation in a wide variety of educational settings. Such discourse facilitation strategies have been recently utilized to design agent support aiming at eliciting student reasoning and engaging students in productive conversational interactions instead of providing content-specific instructional assistance. After highlighting a series of open research issues concerning the effective design and usage of APT agents, we present a prototype web-based conversational agent system developed as a research platform. This system, called MentorChat, was designed deriving from both the established value of APT-based interventions in helping students engage in productive peer interactions and the important role of the teacher in orchestrating collaborative learning activities. The specific system draws away from the typical design of dialogue-based dynamic support systems, tailored to specific learning populations and domains. Specifically, the system enables educators to deploy online dialogue-based activities, asking students to collaborate in order to complete a learning task, as well as configure the domain model of a conversational agent, scaffolding students’ discussions through a series of dynamic APT interventions. Utilizing the easily customizable environment of MentorChat, this dissertation presents six research studies, carried out in two languages (Greek and English), four domains (language learning, multimedia learning, modern history and human-computer interaction) and two education levels (primary and higher education). The studies provide an empirical foundation for an automated, flexible dialogue facilitation approach and contribute to developing an insight into how specific types of APT interventions can be effectively employed in the context of small group collaborative learning activities. Overall, the studies demonstrate that the use of APT agents can be pedagogically beneficial at both individual and group levels by efficiently helping learners to explicitly articulate their reasoning and sustain a productive peer dialogue. Moreover, the students favorably perceive the agent as a group discussion facilitator. Implications for effective design and fine-tuning of agent intervention modes are also discussed, illustrating how specific design choices can drastically improve the pedagogical effectiveness of a conversational agent providing collaborative learning support.

Degree

thesis:*
Grantor dc:publisher
Aristotle University Of Thessaloniki (AUTH)
Year dc:date
2016

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Tegos, Stergios
  • Τέγος, Στέργιος

Subjects

dc:subject × 12

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/38148

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Tegos, Stergios; Τέγος, Στέργιος. Web-based conversational agents for collaborative learning support. Aristotle University Of Thessaloniki (AUTH), 2016. http://hdl.handle.net/10442/hedi/38148