Johan Boye

Plug and Play speech understanding

Plug and Play is an increasingly important concept in system and network architectures. We introduce and describe a spoken language dialogue system architecture which supports Plug and Playable networks of objects in its domain. Each device in the network carries the linguistic and dialogue management information which is pertinent to it and uploads it dynamically to the relevant language processing components in the spoken language interface. We describe the current state of our plug and play demonstrator and discuss theoretical issues that arise from our work. Plug and Play forms a central topic for the DHomme project.

Interaction strategies for an affective conversational agent

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.

Directional types and the annotation method

Abstract A directonal type for a Prolog program expresses certain properties of the operational semantics of the program. This paper shows that the annotation proof method, proposed by Deransart for proving declarative properties of logic programs, is also applicable for proving correctness of directional types. In particular, the sufficient correctness criterion of well-typedness by Bronsard et al., turns out to be a specialization of the annotation method. The comparison shows a general mechanism for construction of similar specializations, which is applied to derive yet another concept of well-typedness. The usefulness of the new correctness criterion is shown on examples of Prolog programs, where the traditional notion of well-typedness is not applicable. We further show that the new well-typing condition can be applied to different execution models. This is illustrated by an example of an execution model where unification is controlled by directional types, and where our new well-typing condition is applied to show the absence of deadlock.

Walk This Way: Spatial Grounding for City Exploration

Recently there has been an interest in spatially aware systems for pedestrian routing and city exploration, due to the proliferation of smartphones with GPS receivers among the general public. Since GPS readings are noisy, giving good and well-timed route instructions to pedestrians is a challenging problem. This paper describes a spoken-dialogue prototype for pedestrian navigation in Stockholm that addresses this problem by using various grounding strategies.

Providing computer game characters with conversational abilities

This paper presents the NICE fairy-tale game system, which enables adults and children to engage in conversation with animated characters in a 3D world. In this paper we argue that spoken dialogue technology have the potential to greatly enrichen the users experience in future computer games. We also present some requirements that have to be fulfilled to successfully integrate spoken dialogue technology with a computer game application. Finally, we briefly describe an implemented system that has provided computer game characters with some conversational abilities that kids have interacted with in studies.

Synthesis of Directionality Information for Functional Logic Programs

Many functional logic programming languages are based on reduction of functional expressions. This feature is also provided by many Prolog systems that offer the facility of calling external functions written in non-logic programming languages. A basic requirement is usually that the arguments of the functions must be ground at invocation time, otherwise an error is reported, or the call is delayed until the arguments are sufficiently instantiated. The drawback of the latter method is twofold: (1) the arguments might never be instantiated, and (2) the dynamic checks made by the delaying mechanism are expensive. This paper presents a method, which for a given program identifies a class of atomic goals for which (1) will not occur. Moreover, we describe a method for transforming a program into an equivalent program, for which dynamic delays are avoided. The static analysis is based on the concept of dependency graphs over an automatically annotated program, a technique originally introduced in connection of attribute grammars.

Persuasive dialogue based on a narrative theory: an ECA implementation

Embodied Conversational Agents (ECA) are poised to constitute a specific category within persuasive systems, in particular through their ability to support affective dialogue. One possible approach consists in using ECA as virtual coaches or personal assistants and to make persuasion part of a dialogue game implementing specific argumentation or negotiation features. In this paper, we explore an alternative framework, which emerges from the long-term development of ECA as “Companions” supporting free conversation with the user, rather than task-oriented dialogue. Our system aims at influencing user attitudes as part of free conversation, albeit on a limited set of topics. We describe the implementation of a Companion ECA to which the user reports on his working day, and which can assess the user’s emotional attitude towards daily events in the office, trying to influence such attitude using affective strategies derived from a narrative model. This discussion is illustrated through examples from a first fully-implemented prototype.

Robust spoken language understanding in a computer game

We present and evaluate a robust method for the interpretation of spoken input to a conversational computer game. The scenario- of the game is that of a player interacting with embodied fairy-tale ...

S-SLD-resolution — An operational semantics for logic programs with external procedures

This paper presents a new operational semantics for logic programs with external procedures, introduced in [BM88]. A new resolution procedure S-SLD-resolution is defined, in which each step of computation is characterized by a goal and a set of equational constraints, whose satisfiability cannot be decided with the information at hand. This approach improves the completeness of the resulting system, since further computation may result in the information needed to solve some earlier unsolved constraints. We also state a sufficient condition to distinguish a class of programs where no unsolved constraints will remain at the end of computation.

Multi-slot semantics for natural-language call routing systems

Statistical classification techniques for natural-language call routing systems have matured to the point where it is possible to distinguish between several hundreds of semantic categories with an accuracy that is sufficient for commercial deployments. For category sets of this size, the problem of maintaining consistency among manually tagged utterances becomes limiting, as lack of consistency in the training data will degrade performance of the classifier. It is thus essential that the set of categories be structured in a way that alleviates this problem, and enables consistency to be preserved as the domain keeps changing. In this paper, we describe our experiences of using a two-level multi-slot semantics as a way of meeting this problem. Furthermore, we explore the ramifications of the approach with respect to classification, evaluation and dialogue design for call routing systems.