TY - GEN
T1 - Socially Aware Interactions
T2 - 5th International Workshop on Chatbot Research and Design, CONVERSATIONS 2021
AU - Lobo, Inês
AU - Rato, Diogo
AU - Prada, Rui
AU - Dignum, Frank
PY - 2022
Y1 - 2022
N2 - In this paper, we present a prototype of a human-agent dialogue system, in which the scenarios are easy-to-author, as in tree-based dialogue tools. These, however, only allow for scripted and restricted dialogues. For this reason, we focused on developing a flexible and robust deliberation mechanism as well, based on the Cognitive Social Frames model and the theory of social practices, so that the conversational agent could provide acceptable responses according to different social contexts. Having access to sequences of frames containing small dialogue trees, the agent activates the most salient frame to reply appropriately to the user’s input. As a proof of concept, we designed a medical diagnosis scenario between a doctor and a patient in which the agent could play both roles given different settings of the scenario. In this prototype, the user had to choose from a limited set of alternatives, based on the current context, in order to respond to the agent; however, in the future, we intend to allow users to write freely, expecting to be able to map their utterances to the appropriate context.
AB - In this paper, we present a prototype of a human-agent dialogue system, in which the scenarios are easy-to-author, as in tree-based dialogue tools. These, however, only allow for scripted and restricted dialogues. For this reason, we focused on developing a flexible and robust deliberation mechanism as well, based on the Cognitive Social Frames model and the theory of social practices, so that the conversational agent could provide acceptable responses according to different social contexts. Having access to sequences of frames containing small dialogue trees, the agent activates the most salient frame to reply appropriately to the user’s input. As a proof of concept, we designed a medical diagnosis scenario between a doctor and a patient in which the agent could play both roles given different settings of the scenario. In this prototype, the user had to choose from a limited set of alternatives, based on the current context, in order to respond to the agent; however, in the future, we intend to allow users to write freely, expecting to be able to map their utterances to the appropriate context.
UR - http://www.scopus.com/inward/record.url?scp=85124659373&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-94890-0_8
DO - 10.1007/978-3-030-94890-0_8
M3 - Conference contribution
SN - 9783030948894
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 124
EP - 140
BT - Chatbot Research and Design - 5th International Workshop, CONVERSATIONS 2021, Revised Selected Papers
A2 - Følstad, A.
A2 - Araujo, T.
A2 - Papadopoulos, S.
A2 - Law, E.L.-C.
A2 - Luger, E.
A2 - Goodwin, M.
A2 - Brandtzaeg, P.B.
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 23 November 2021 through 24 November 2021
ER -