Interaction Strategies for an Affective Conversational Agent

Abstract

The development of embodied conversational agents (ECA) as companions brings several challenges for both affective and conversational dialogue. These include challenges in generating appropriate affective responses, selecting the overall shape of the dialogue, providing prompt system response times, and handling interruptions. We present an implementation of such a companion showing the development of individual modules that attempt to address these challenges. Further, to resolve resulting conflicts, we present encompassing interaction strategies that attempt to balance the competing requirements along with dialogues from our working prototype to illustrate these interaction strategies in operation. Finally, we provide the results of an evaluation of the companion using an evaluation methodology created for conversational dialogue and including analysis using appropriateness annotation.

Toward a memory assistant companion for the individuals with mild memory impairment

“…Leite et al [9] investigated a robot cat showing empathic behaviors (voice, facial expressions, and head movements) on the players of chess game. Smith et al [10] proposed the integration of affective dialogue with a deliberative architecture. These studies showed that the display of empathic behaviors can usually make the conversational artifacts better accepted by users, which is a requirement of artificial companions.…”

Proceedings of the Workshop on NLG for Human–Robot Interaction

“…speaking or planning an utterance. As in other situated communication tasks (Koller et al, 2010;Smith et al, 2011) the timing of the robot's utterances is important. For fluent interactions the robot needs to monitor the human's actions and changes in the environment and react to them in a timely manner, potentially by interrupting itself or modifying an utterance mid-stream (Clark and Krych, 2004).…”

Grounding the detection of the user’s likes and dislikes on the topic structure of human-agent interactions

“…Such a trend is due to the increasing number of web video reviews available under platforms such as Youtube. Beside the existing literature on multimodal sentiment analysis, we have found only two studies, [4] and [5] , using a module for the detection of verbal-centred sentiment-related phenomena in human-agent interactions: the studies describe an implementation of "How was your day" Companions integrating a Sentiment Analysis module grounded on the theoretical model provided by [6] and initially designed for the analysis of non-conversational texts. These studies have the merit to show the importance of the sentiment-related detection during a conversation between a virtual agent and a human user.…”

“…These studies have the merit to show the importance of the sentiment-related detection during a conversation between a virtual agent and a human user. However, as explained in Clavel et al [7] , the systems described by [4] and [5] are learned with non-conversational texts and not properly designed to be fit for the conversational context and human-agent interaction. Indeed, designing a system able to detect verbal expressions of sentiment during face-to-face and speech-to-speech interactions between a human user and a virtual agent requires to deal with various issues.…”

“…Example of the Semaine Corpus, Poppy, session 21 5. Example of the Semaine Corpus, Poppy, session 26 6.…”