Human-Machine Dialogue as a Stochastic Game

Barlier, Merwan; Pérolat, Julien; Laroche, Romain; Pietquin, Olivier

doi:10.18653/v1/w15-4602

Cited by 15 publications

(10 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…It would also be interesting to introduce modifications of the algorithms to allow more general value function representations than linear parametrization. Co-adaptation between humans and CA [32] should also be studied. Finally, this approach will be tested on real human-human dialogues.…”

Section: Resultsmentioning

confidence: 99%

A Stochastic Model for Computer-Aided Human-Human Dialogue

Barlier¹,

Laroche²,

Pietquin³

2016

Interspeech 2016

Self Cite

View full text Add to dashboard Cite

In this paper we introduce a novel model for computer-aided human-human dialogue. In this context, the computer aims at improving the outcome of a human-human task-oriented dialogue by intervening during the course of the interaction. While dialogue state and topic tracking in human-human dialogue have already been studied, few work has been devoted to the sequential part of the problem, where the impact of the system's actions on the future of the conversation is taken into account. This paper addresses this issue by first modelling human-human dialogue as a Markov Reward Process. The task of purposely taking part into the conversation is then optimised within the Linearly Solvable Markov Decision Process framework. Utterances of the Conversational Agent are seen as perturbations in this process, which aim at satisfying the user's long-term goals while keeping the conversation natural. Finally, results obtained by simulation suggest that such an approach is suitable for computer-aided human-human dialogue and is a first step towards three-party dialogue.

show abstract

Section: Resultsmentioning

confidence: 99%

A Stochastic Model for Computer-Aided Human-Human Dialogue

Barlier¹,

Laroche²,

Pietquin³

2016

Interspeech 2016

Self Cite

View full text Add to dashboard Cite

show abstract

“…On top of all that, the conversation situation implies that the system might be able to participate in the argumentation about the discussed topics. The argumentative capacity, despite recent very promising works based on the use of game theory [2,13], remains very difficult to implement in the current architecture. This is a point that our endeavours to convert a goal-oriented system to a conversational interface have not yet been able to address appropriately, as discussed in the next section.…”

Section: Comparison Of Goal-oriented Vs Conversational Agentsmentioning

confidence: 99%

Lilia, A Showcase for Fast Bootstrap of Conversation-Like Dialogues Based on a Goal-Oriented System

Riou

Jabaian

Huet

et al. 2019

Statistical Language and Speech Processing

View full text Add to dashboard Cite

Recently many works have proposed to cast human-machine interaction in a sentence generation scheme. Neural networks models can learn how to generate a probable sentence based on the user's statement along with a partial view of the dialogue history. While appealing to some extent, these approaches require huge training sets of generalpurpose data and lack a principled way to intertwine language generation with information retrieval from back-end resources to fuel the dialogue with actualised and precise knowledge. As a practical alternative, in this paper, we present Lilia, a showcase for fast bootstrap of conversation-like dialogues based on a goal-oriented system. First, a comparison of goaloriented and conversational system features is led, then a conversion process is described for the fast bootstrap of a new system, finalised with an on-line training of the system's main components. Lilia is dedicated to a chitchat task, where speakers exchange viewpoints on a displayed image while trying collaboratively to derive its author's intention. Evaluations with user trials showed its efficiency in a realistic setup.

show abstract

“…A medida que mejora la comprensión del lenguaje y la tecnología de generación automática de diálogos, también aumenta el interés en la construcción de sistemas de conversación de usuario, que pueden ser utilizados para una variedad de aplicaciones tales como planificación de viajes, sistemas tutoriales o soporte técnico basado en chat. La descripción de algunos de estos sistemas desarrollados en los últimos años se puede encontrar en [1][2][3][4][5][6][7][8].…”

Section: Introductionunclassified

Identificación de etiquetas semánticas para su uso en diálogos

Vázquez¹,

Pinto²,

Vilariño³

2018

RCS

View full text Add to dashboard Cite

Resumen. En la interacción humano-robot la generación automática de diálogos es una parte importante. Los diálogos generados deben garantizar una conversación coherente entre el humano y el robot, se espera que la interacción sea lo más natural y eficaz que se pueda. En este trabajo se plantea el uso de entidades semánticas como elementos básicos que posteriormente serán utilizados para la realización de un módulo de comprensión del lenguaje como parte de un sistema de diálogos. Para poder identificar las entidades semánticas en los diálogos se evaluaron dos herramientas para reconocer entidades nombradas (NER), que utilizan como núcleo un clasificador basado en Campos Aleatorios Condicionales. Los resultados que se obtienen en este trabajo permiten afirmar que estos dos NER tienen buenos resultados en el reconocimiento de las entidades semánticas (tokens) pues de manera global superan el 84% de exactitud. El corpus utilizado en este trabajo es el corpus español DIHANA, el cual está compuesto de diálogos sobre un sistema de información de consultas telefónicas sobre horarios y precios de trenes de largo recorrido.Palabras clave: sistema de diálogos, entidad semántica, reconocedor de entidad nombrada.Abstract. In the human-robot interaction the automatic generation of dialogues is an important part, the generated dialogues must guarantee a coherent conversation between the human and the robot, the interaction is expected to be as natural and effective as possible. In this paper, the use of semantic entities as basic elements is proposed, which will later be used for the realization of a language understanding module as part of a dialog system. To identify the semantic entities in the dialogues, two tools for recognizing named entities (NER) were evaluated, using as a nucleus a classifier based on Conditional Random Fields. The results obtained in this work allow us to affirm that these two NERs have satisfactory results in the recognition of semantic entities (tokens) because, overall, they exceed 84% accuracy. The corpus used in this 99

show abstract

Human-Machine Dialogue as a Stochastic Game

Cited by 15 publications

References 30 publications

A Stochastic Model for Computer-Aided Human-Human Dialogue

A Stochastic Model for Computer-Aided Human-Human Dialogue

Lilia, A Showcase for Fast Bootstrap of Conversation-Like Dialogues Based on a Goal-Oriented System

Identificación de etiquetas semánticas para su uso en diálogos

Contact Info

Product

Resources

About