Imitation Learning Applied to Embodied Conversational Agents

Piot Bilal, Olivier Pietquin, Matthieu Geist
; Proceedings of The 4th Workshop on Machine Learning for Interactive Systems at ICML 2015, PMLR 43:1-5, 2015.

Abstract

Embodied Conversational Agents (ECAs) are emerging as a key component to allow human interact with machines. Applications are numerous and ECAs can reduce the aversion to interact with a machine by providing user-friendly interfaces. Yet, ECAs are still unable to produce social signals appropriately during their interaction with humans, which tends to make the interaction less instinctive. Especially, very little attention has been paid to the use of laughter in human-avatar interactions despite the crucial role played by laughter in human-human interaction. In this paper, methods for predicting when and how to laugh during an interaction for an ECA are proposed. Different Imitation Learning (also known as Apprenticeship Learning) algorithms are used in this purpose and a regularized classification algorithm is shown to produce good behavior on real data.

Cite this Paper


BibTeX
@InProceedings{pmlr-v43-piot15, title = {Imitation Learning Applied to Embodied Conversational Agents}, author = {Piot Bilal and Olivier Pietquin and Matthieu Geist}, booktitle = {Proceedings of The 4th Workshop on Machine Learning for Interactive Systems at ICML 2015}, pages = {1--5}, year = {2015}, editor = {Heriberto Cuayáhuitl and Nina Dethlefs and Lutz Frommberger and Martijn Van Otterlo and Olivier Pietquin}, volume = {43}, series = {Proceedings of Machine Learning Research}, address = {Lille, France}, month = {11 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v43/piot15.pdf}, url = {http://proceedings.mlr.press/v43/piot15.html}, abstract = {Embodied Conversational Agents (ECAs) are emerging as a key component to allow human interact with machines. Applications are numerous and ECAs can reduce the aversion to interact with a machine by providing user-friendly interfaces. Yet, ECAs are still unable to produce social signals appropriately during their interaction with humans, which tends to make the interaction less instinctive. Especially, very little attention has been paid to the use of laughter in human-avatar interactions despite the crucial role played by laughter in human-human interaction. In this paper, methods for predicting when and how to laugh during an interaction for an ECA are proposed. Different Imitation Learning (also known as Apprenticeship Learning) algorithms are used in this purpose and a regularized classification algorithm is shown to produce good behavior on real data.} }
Endnote
%0 Conference Paper %T Imitation Learning Applied to Embodied Conversational Agents %A Piot Bilal %A Olivier Pietquin %A Matthieu Geist %B Proceedings of The 4th Workshop on Machine Learning for Interactive Systems at ICML 2015 %C Proceedings of Machine Learning Research %D 2015 %E Heriberto Cuayáhuitl %E Nina Dethlefs %E Lutz Frommberger %E Martijn Van Otterlo %E Olivier Pietquin %F pmlr-v43-piot15 %I PMLR %J Proceedings of Machine Learning Research %P 1--5 %U http://proceedings.mlr.press %V 43 %W PMLR %X Embodied Conversational Agents (ECAs) are emerging as a key component to allow human interact with machines. Applications are numerous and ECAs can reduce the aversion to interact with a machine by providing user-friendly interfaces. Yet, ECAs are still unable to produce social signals appropriately during their interaction with humans, which tends to make the interaction less instinctive. Especially, very little attention has been paid to the use of laughter in human-avatar interactions despite the crucial role played by laughter in human-human interaction. In this paper, methods for predicting when and how to laugh during an interaction for an ECA are proposed. Different Imitation Learning (also known as Apprenticeship Learning) algorithms are used in this purpose and a regularized classification algorithm is shown to produce good behavior on real data.
RIS
TY - CPAPER TI - Imitation Learning Applied to Embodied Conversational Agents AU - Piot Bilal AU - Olivier Pietquin AU - Matthieu Geist BT - Proceedings of The 4th Workshop on Machine Learning for Interactive Systems at ICML 2015 PY - 2015/06/18 DA - 2015/06/18 ED - Heriberto Cuayáhuitl ED - Nina Dethlefs ED - Lutz Frommberger ED - Martijn Van Otterlo ED - Olivier Pietquin ID - pmlr-v43-piot15 PB - PMLR SP - 1 DP - PMLR EP - 5 L1 - http://proceedings.mlr.press/v43/piot15.pdf UR - http://proceedings.mlr.press/v43/piot15.html AB - Embodied Conversational Agents (ECAs) are emerging as a key component to allow human interact with machines. Applications are numerous and ECAs can reduce the aversion to interact with a machine by providing user-friendly interfaces. Yet, ECAs are still unable to produce social signals appropriately during their interaction with humans, which tends to make the interaction less instinctive. Especially, very little attention has been paid to the use of laughter in human-avatar interactions despite the crucial role played by laughter in human-human interaction. In this paper, methods for predicting when and how to laugh during an interaction for an ECA are proposed. Different Imitation Learning (also known as Apprenticeship Learning) algorithms are used in this purpose and a regularized classification algorithm is shown to produce good behavior on real data. ER -
APA
Bilal, P., Pietquin, O. & Geist, M.. (2015). Imitation Learning Applied to Embodied Conversational Agents. Proceedings of The 4th Workshop on Machine Learning for Interactive Systems at ICML 2015, in PMLR 43:1-5

Related Material