Planning and Logic Programming for Dialog Management

This paper reports about our research to design and implement a spoken language dialogue system (SLDS) that can be con gured easily and rapidly to di erent applications. In contrast to nite-state models for information dialogues, we outline a different knowledge-based approach capable to handle dialogues of greater complexity than those following a nite-state argumentation structure. We show how knowledge about the application domain, the dialogue participants, syntax and semantics can be represented in a uni ed framework based on Description Logics (DL) as modelling formalism. On the basis of the knowledge in the domain model, inferences are performed employing a partial rst order logic for the interpretation of user utterances and the planning of system responses.

Proceedings of the 1st SIGdial workshop on Discourse and dialogue -, 2000

Intelligent dialogue systems must be able to respond properly to a variety of requests involving knowledge of the dialogue, the task at hand, and the domain. This requires advanced knowledge reasoning performed by various processing modules. We argue that it is important to understand the nature of the various reasoning mechanisms involved and to separate not only, for instance, interpretation, generation, and dialogue management but also domain knowledge and task reasoning. This facilitates portability of the dialogue system to new domains and makes it easier to enhance its capabilities. In this paper we will focus on the dialogue and domain knowledge reasoning components and show h o w they can cooperate to achieve natural interaction.

Proceedings of the 3rd Language and Technology …, 2007

2009 IEEE International Conference on Multimedia and Expo, 2009

ABSTRACT The importance of dialog management systems has increased in recent years. Dialog systems are created for domain specific applications, so that a high demand for a flexible dialog system framework arises. There are two basic approaches for dialog management systems: a rule-based approach and a statistic approach. In this paper, we combine both methods and form a hybrid dialog management system in a scalable agent based framework. For deciding of the next dialog step, two independent systems are used: the Java Rule Engine (JESS) as expert system for rule-based solutions, and the Partially Observable Markov Decision Process (POMDP) as model-based solution for more complex dialog sequences. Using a speech recognizer and text-to-speech systems, the human can be guided through a dialog with approximately ten steps.

1997

In order to participate in dialogues, an agent needs the capability of modeling mental states. Specifically, it is necessary to represent the agent attitudes (beliefs, intentions, and objectives), world knowledge, and temporal, reasoning and behavior rules. In this paper, we propose a logic programming framework that allows the representation of agent models. Agents are defined as logic programs that are extended with explicit negation.

Proceedings, 2000

Simple question answering systems are no longer enough to answer the complex questions asked by users, and increasingly there is the need to enter into dialogue in order to re ne or reiterate user requests. Complex Dialogue Management (DM) components have been developed ...

In interactive systems, the ~dialogue manager' is a central component, whose primary task is to decide what to do next in a dialogue. This decision involves taking a variety of kinds of context information into account, with different logical properties, used in different types of reasoning, and therefore with different demands on representation. In this paper we discuss the representation of the context information needed to support intelligent dialogue management in interactive speech systems. We argue that simple types of context information should be represented in a way that optimally supports efficient simple forms of reasoning, while other types of context information require sophisticated logics for articulate representation and full-blown reasoning. Using different types of representation for different types of context information creates a problem in situations where a type of context information, that is adequately represented in a very simple form most of the time, exceptionally requires articulate representation and sophisticated reasoning. We indicate how this problem can be resolved, and how a dialogue manager can use several forms of context representation and reasoning for generating dialogue acts, by clearly distinguishing the different ways in which dialogue acts may be constructed and the conditions under which a particular construction is to be used.

2011

AbstrAct: This paper focuses our agent-based approach to dialogue management equipped with a deliberation mechanism further extensible with additional features. The aim of this study has been to create a manager which can exhibit complex behaviour from two points of view – the development and runtime ones. To support the first case, the domain developer needs to provide a set of plans which tell the agent how to cope with particular tasks. Then during the runtime, the agent processes these plans to manage the dialogu e in accordance with an optimization criterion specified. This paper summarizes our effort and gives an overview about our architecture.

2005

Modelling rational dialogues on the basis of FiniteState Automata has shown to be of limited use when natural speech behaviour and dynamic adaptation to new domains are an issue. To enable such a flexible, robust and generic dialog ue management, we rely on an approach characterized by a logica lly precise modelling of the linguistic and application aspect s of the domain in question and a dialogue manager deriving pragmati csdriven discourse strategies from reasoning on evolving bel i f structures, both of the user and the system. For the linguist ic part, we claim that a “pragmatics-first” view on rational interact ion provides an appropriate framework for flexible and scalable dialogue modelling. In particular, the plan-based approach offers the means to conduct taskor goal-oriented dialogues which aim at accomplishing concrete tasks. It enables cooperative re sponse behaviour and the ability for negotiation. For the reasoning part, i.e. knowledge representation and inference for...

Naacl, 2001

Developing dialogue systems is a complex process. In particular, designing efficient dialogue management strategies is often difficult as there are no precise guidelines to develop them and no sure test to validate them. Several suggestions have been made recently to use reinforcement learning to search for the optimal management strategy for specific dialogue situations. These approaches have produced interesting results, including applications involving real world dialogue systems. However, reinforcement learning suffers from the fact that it is state based. In other words, the optimal strategy is expressed as a decision table specifying which action to take in each specific state. It is therefore difficult to see whether there is any generality across states. This limits the analysis of the optimal strategy and its potential for re-use in other dialogue situations. In this paper we tackle this problem by learning rules that generalize the state-based strategy. These rules are more readable than the underlying strategy and therefore easier to explain and re-use. We also investigate the capability of these rules in directing the search for the optimal strategy by looking for generalization whilst the search proceeds.