Núria Castell

A portable method for acquiring information extraction patterns without annotated corpora

The main issue when building Information Extraction (IE) systems is how to obtain the knowledge needed to identify relevant information in a document. Most approaches require expert human intervention in many steps of the acquisition process. In this paper we describe ESSENCE, a new method for acquiring IE patterns that significantly reduces the need for human intervention. The method is based on ELA, a specifically designed learning algorithm for acquiring IE patterns without tagged examples. The distinctive features of ESSENCE and ELA are that (1) they permit the automatic acquisition of IE patterns from unrestricted and untagged text representative of the domain, due to (2) their ability to identify regularities around semantically relevant concept-words for the IE task by (3) using non-domain-specific lexical knowledge tools such as WordNet, and (4) restricting the human intervention to defining the task, and validating and typifying the set of IE patterns obtained. Since ESSENCE does not require a corpus annotated with the type of information to be extracted and it uses a general purpose ontology and widely applied syntactic tools, it reduces the expert effort required to build an IE system and therefore also reduces the effort of porting the method to any domain. The results of the application of ESSENCE to the acquisition of IE patterns in an MUC-like task are shown.

Combining Knowledge and Metrics to Control Software Quality Factors

The LESD project (Linguistic Engineering for Software Development) aimed to develop computing tools for analysis and reasoning on functional or preliminary specifications of aerospace software written in English. These tools help to control the quality of software written during the first stage: specification. The factors considered relevant to the quality of specifications in the LESD project are: traceability, modifiability, completeness, consistency, and verifiability. This paper deals with completeness and modifiability. In the case of completeness we present a symbolic approach to control this factor, using a Knowledge Base. Checks are based on metarequirements that try to ensure structural completeness. The concept of modifiability is based on the level of interconnection between the requirements of the specifications. Two metrics have been defined in order to measure global and local levels of interconnection.

Filtering Software Specifications Written in Natural Language

The specification phase is one of the most important and least supported part of the software development process. We have conceived SAREL (Assistance System for Writing Software Specification in Natural Language) as a tool to improve the specification phase. SAREL is a continuation of a program of research and development called LESD (Linguistic Engineering for Software Design). The purpose of SAREL is to assist engineers in the creation of software specifications written in natural language. It is divided into three modules: the first one controls the requirement according to the writing norms, the second one obtains a conceptual representation using the Knowledge Base, and the third one carries out a series of optional analyses taking into account the following software quality properties: consistency, completeness, traceability, verifiability and modifiability. Once a requirement has been labeled as correct, its conceptual representation is added to the Requirements Base.

Linguistic and Logical Tools for an Advanced Interactive Speech System in Spanish

This paper focuses on the increasing need for a more natural and sophisticated human-machine interaction (HMI). The research here presented shows work on the development of a restricted-domain spontaneous speech dialogue system in Spanish. This human-machine interface is oriented towards a semantically restricted domain: Spanish railway information. The paper focuses on the description of the understanding module, which performs the language processing once the dialogue moves have been recognised and transcribed into text. Following the morphological, syntactic and semantic analysis, the module generates a structured representation with the content of the users intervention. This representation is passed on to the dialogue manager, which generates the systems answer. The dialogue manager keeps the dialogue history and decides what the reaction of the system should be, expressed by a new structured representation. This is sent to the natural language generator, which then builds the sentence to be synthesised.

Architecture and dialogue design for a voice operated information system

In this paper we present a real automatic meteorological information system that, not only provides friendly voice access to real-time data coming from automatic sensors, but also establishes an automatic warning service on the weather. It aims to extend the availability, personalization and friendliness of the meteorological information by means of a reusable easy-to-use friendly oral natural language interface. This interface takes advantage of the improvements in speech processing, dialogue handling and the great growth of mobile telephony. After the description of the functionalities of the system and its architecture, we present in detail the features of the dialogue manager. The main goals we have considered are: to provide the right information and to design a friendly interface.

Dialogue management in an automatic meteorological information system

In this paper we present a real automatic meteorological information system that, not only provides friendly voice access to real-time data coming from automatic sensors, but also establishes an automatic warning service on the weather. It aims to extend the availability, personalization and friendliness of the meteorological information by means of a reusable easy-to-use friendly oral natural language interface. This interface takes advantage of the improvements in speech processing, dialogue handling and the great growth of mobile telephony. After the description of the functionalities of the system and its architecture, we present in detail the features of the dialogue manager. The main goals we have considered are: to provide the right information, to design a friendly interface, and to help the user never getting lost during the dialogue.

Strategies to Overcome Problematic Input in a Spanish Dialogue System

This paper focuses on the strategies adopted to tackle problematic input and ease communication between modules in a Spanish railway information dialogue system for spontaneous speech. The paper describes the design and tuning considerations followed by the understanding module, both from a language processing and semantic information extraction point of view. Such strategies aim to handle the problematic input received from the speech recogniser, which is due to spontaneous speech as well as recognition errors.

Knowledge management in the Sarel system

The specification phase is one of the most important and least supported parts of the software development process. SAREL (Assistance System for Writing Software Specifications in Natural Language) has been conceived as a knowledge based tool to improve the specification phase. The purpose of SAREL is to assist engineers in the creation of software specifications written in natural language (NL). The aim of the paper is to present the creation process of the knowledge base and the requirements base corresponding to a software specification written in NL. This process uses some linguistic engineering tools developed in our department designed to process documents that contain preliminary software specifications. These documents are divided into several parts. We can distinguish the Introduction and the Overall Description as parts that should be used by the knowledge base construction. Up to a certain point, these two parts contain all the background information needed to understand the problem as a whole. The information contained in the specific requirements section corresponds to the information represented in the requirements base.