Paul Piwek

A flexible pragmatics-driven language generator for animated agents

This paper describes the NECA MNLG; a fully implemented Multimodal Natural Language Generation module. The MNLG is deployed as part of the NECA system which generates dialogues between animated agents. The generation module supports the seamless integration of full grammar rules, templates and canned text. The generator takes input which allows for the specification of syntactic, semantic and pragmatic constraints on the output.

Incremental conceptualization for language production

For the past ten years or more, most work in the field of Natural Language Generation (NLG) has shied away from considerations regarding the processes underlying human language production. Rather, the focus has been on systems that automatically produce language—usually text—from non-linguistic representations, with the main objective being generation of a text that faithfully captures the meaning of those nonlinguistic representations (see, e.g., Reiter and Dale’s 2000 textbook on NLG). There is, however, also a different take on NLG “as not just competent performance by a computer but the development of a computational theory of the human capacity for language and processes that engage it” (McDonald 1987, page 642). Guhe’s research monograph, based on his 2003 Ph.D. thesis, is firmly situated in the latter tradition. One of his main goals is to work out a computational architecture for Levelt’s (1989) psycholinguistically motivated model of language production. According to Levelt’s model, speaking involves three main activities: conceptualizing (deciding what to say), formulating (deciding how to say it), and articulating (saying it). Guhe’s book focuses on the mental activity of conceptualizing. Conceptualizing is a recalcitrant object of study, partly because of the problem of the “initial spark”; the decision to say something appears to be the result of volitional conscious decisions, which largely elude scientific study. Guhe avoids this problem by investigating conceptualization in settings where the main intention is already fixed: a speaker witnesses several events unfold and is instructed to describe what happens (while it happens). The research challenge then is to figure out how “subintentions” for individual speech acts come about. The benefit of using an on-line generation setting is that it provides information on both what a speaker says at a given point in time and what is being reported, that is, the data that drive the speaker’s utterances. The book consists of the usual preface and introduction, followed by four parts (A, B, C, and Results), a list of the book’s theses, and an appendix that includes, among other things, a glossary, bibliography, name index, and subject index. Part A of the book is titled “Conceptualization.” It starts with an introduction to the field of language production, with particular reference to Levelt’s (1989) model. The notion of conceptualization as a “quasi-module,” partly using Fodor’s (1983) criteria, is presented and four subtasks of conceptualization are discussed:A process for producing high-quality alpha -olefin polymers with a higher yield in a long lasting stabilized manner even in the case of gas phase polymerization is provided. In the production process of alpha -olefin polymers which comprises reacting a trivalent metal halide with a divalent metal compound to obtain a solid product (I); reacting this product with at least one electron donor (ED) and at least one electron acceptor (EA), once to 10 times, and at that time, using TiCl4 at least once as the (EA) to obtain a solid product (II); combining this product with an organoaluminum compound (OAl) and an (ED) (these three being referred to as catalyst components), the improvement which comprises subjecting a part or the whole of the catalyst components to polymerization treatment with a small amount of an alpha -olefin, at least in the coexistence of the solid product (II) and (OAl), in the combination of the catalyst components to obtain a preactivated catalyst, and subjecting alpha -olefin(s) to gas phase polymerization or bulk or slurry polymerization followed by gas phase polymerization, using the catalyst.

Making Tacit Requirements Explicit

The importance of tacit knowledge in Requirements Engineering (RE) is widely acknowledged. While valuable work has developed techniques to expose sources of tacit knowledge during requirements elicitation, such techniques are not universally applied and tacit knowledge, continues to negatively affect the quality of the requirements. In this position paper we present a brief review and interpretation of the literature on tacit knowledge that, we believe, is useful for RE. We describe a number of techniques that offer analysts the means to reason about the effect of tacit knowledge and improve the quality of requirements and their management.

Generating monologue and dialogue to present personalised medical information to patients

Medical information is notoriously difficult to convey to patients because the content is complex, emotionally sensitive, and hard to explain without recourse to technical terms. We describe a pilot system for communicating the contents of electronic health records (EHRs) to patients. It generates two alternative presentations, which we have compared in a preliminary evaluation study: the first takes the form of a monologue, which elaborates the information taken from the patients EHR by adding explanations of some concepts and procedures; the second takes the form of a scripted dialogue, in which the content is recast as a series of questions, answers and statements assigned to two characters in the dialogue, a senior and a junior nurse. Our discourse planning method designs these presentations in tandem, first producing a monologue plan which is then elaborated into a dialogue plan.

Unpacking Tacit Knowledge for Requirements Engineering

The use of tacit knowledge is a common feature in everyday communication. It allows people to communicate effectively without forcing them to make everything tediously and painstakingly explicit, provided they all share a common understanding of whatever is not made explicit. If this latter criterion does not hold, confusion and misunderstanding will ensue. Tacit knowledge is also commonplace in requirements where it also affords economy of expression. However, the use of tacit knowledge also suffers from the same risk of misunder-standing, with the associated problems of anticipating where it has the potential for confusion, and of unraveling where it has played an actual role in misunder-standing. Thus the effective communication of requirements knowledge (whether verbally, through a document or some other medium) requires an understanding of what knowledge is and isn’t (necessarily) held in common. This is very hard to get right as people from different professional and cultural backgrounds are typically involved. At its worst, tacit requirements knowledge may lead to software that fails to satisfy the customer’s requirements. In this chapter we review the diverse views of tacit knowledge discussed in the literature from a wide range of disci-plines, reflect on their commonalities and differences, and propose a conceptual framework for requirements engineering that characterizes the different facets of tacit knowledge that distinguish the different views. We then identify methodolog-ical and technical challenges for future research on the role of tacit knowledge in requirements engineering.

Generating Multimedia Presentations from Plain Text to Screen Play

In many Natural Language Generation (NLG) applications, the output is limited to plain text – i.e., a string of words with punctuation and paragraph breaks, but no indications for layout, or pictures, or dialogue. In several projects, we have begun to explore NLG applications in which these extra media are brought into play. This paper gives an informal account of what we have learned. For coherence, we focus on the domain of patient information leaflets, and follow an example in which the same content is expressed first in plain text, then in formatted text, then in text with pictures, and finally in a dialogue script that can be performed by two animated agents. We show how the same meaning can be mapped to realisation patterns in different media, and how the expanded options for expressing meaning are related to the perceived style and tone of the presentation. Throughout, we stress that the extra media are not simple added to plain text, but integrated with it: thus the use of formatting, or pictures, or dialogue, may require radical rewording of the text itself.

On Presuppositions in Requirements

Tacit knowledge in requirements documents can lead to miscommunication between software engineers and

Politeness and Bias in Dialogue Summarization: Two Exploratory Studies

In this chapter, two empirical pilot studies on the role of politeness in dialogue summarization are described. In these studies, a collection of four dialogues was used. Each dialogue was automatically generated by the NECA system and the politeness of the dialogue participants was systematically manipulated. Subjects were divided into groups who had to summarize the dialogues from a particular dialogue participant’s point of view or the point of view of an impartial observer. In the first study, there were no other constraints. In the second study, the summarizers were restricted to summaries whose length did not exceed 10% of the number of words in the dialogue that was being summarized.

Natural language processing in clime, a multilingual legal advisory system†

This paper describes CLIME, a web-based legal advisory system with a multilingual natural language interface. CLIME is a ‘proof-of-concept’ system which answers queries relating to ship-building and ship-operating regulations. Its core knowledge source is a set of such regulations encoded as a conceptual domain model and a set of formalised legal inference rules. The system supports retrieval of regulations via the conceptual model, and assessment of the legality of a situation or activity on a ship according to the legal inference rules. The focus of this paper is on the natural language aspects of the system, which help the user to construct semantically complex queries using WYSIWYM technology, allow the system to produce extended and cohesive responses and explanations, and support the whole interaction through a hybrid synchronous/asynchronous dialogue structure. Multilinguality (English and French) is viewed simply as interface localisation: the core representations are language-neutral, and the system can present extended or local interactions in either language at any time. The development of CLIME featured a high degree of client involvement, and the specification, implementation and evaluation of natural language components in this context are also discussed.

Predicting the Understandability of OWL Inferences

In this paper, we describe a method for predicting the understandability level of inferences with OWL. Specifically, we present a probabilistic model for measuring the understandability of a multiple-step inference based on the measurement of the understandability of individual inference steps. We also present an evaluation study which confirms that our model works relatively well for two-step inferences with OWL. This model has been applied in our research on generating accessible explanations for an entailment of OWL ontologies, to determine the most understandable inference among alternatives, from which the final explanation is generated.