Andrea Setzer

Building a semantically annotated corpus of clinical texts

In this paper, we describe the construction of a semantically annotated corpus of clinical texts for use in the development and evaluation of systems for automatically extracting clinically significant information from the textual component of patient records. The paper details the sampling of textual material from a collection of 20,000 cancer patient records, the development of a semantic annotation scheme, the annotation methodology, the distribution of annotations in the final corpus, and the use of the corpus for development of an adaptive information extraction system. The resulting corpus is the most richly semantically annotated resource for clinical text processing built to date, whose value has been demonstrated through its use in developing an effective information extraction system. The detailed presentation of our corpus construction and annotation methodology will be of value to others seeking to build high-quality semantically annotated corpora in biomedical domains.

Multimodal Generation in the COMIC Dialogue System

We describe how context-sensitive, user-tailored output is specified and produced in the COMIC multimodal dialogue system. At the conference, we will demonstrate the user-adapted features of the dialogue manager and text planner.

The Role of Inference in the Temporal Annotation and Analysis of Text

In this paper we argue for the importance of doing inference over the information expressed by the annotations of temporally annotated corpora. We describe the process of inferential closure which can be applied to determine the full temporal content that follows from an annotation. We illustrate the importance of temporal inference and temporal closure in relation to three tasks, which are: (a) the comparison of different temporal annotations, (b) facilitating the manual annotation process needed to create temporally annotated corpora and (c) empirical investigations done over temporally annotated data.

Task-Oriented Extraction of Temporal Information: The Case of Clinical Narratives

Most recent work on temporal relation extraction from text has addressed text drawn from the newswire domain and has attempted to extract all temporal relational information, as specified by proposed temporal annotation schemes such as TimeML. In this paper we explore the task of extracting restricted amounts of temporal information in support of an information extraction application in the medical domain, specifically that of extracting information about times of clinical investigations (X-rays, ultrasounds, etc.) from clinic letters. We describe the task, the corpus and evaluation data we have assembled, a baseline algorithm for extracting temporal relations between temporal expressions and clinical investigation events, and present evaluation results for the algorithm. Overall scores of precision 73.83% and recall 58.70% are promising for a simple baseline approach and suggest that extracting only a restricted subset of the temporal information available in a text may be a sensible way to proceed in the context of specific applications

A pilot study on annotating temporal relations in text

We describe a pilot study in which a scheme for annotating events and temporal relations between events in text is applied to a small corpus of newswire texts. High levels of agreement between human annotators are shown to be difficult to achieve and we investigate where discrepancies occur and how these might be addressed.

Multilingual authoring: the NAMIC approach

With increasing amounts of electronic information available, and the increase in the variety of languages used to produce documents of the same type, the problem of how to manage similar documents in different languages arises. This paper proposes an approach to processing/structuring text so that Multilingual Authoring (creating hypertext links) can be effectively carried out. This work, funded by the European Union, is applied to the Multilingual Authoring of news agency text. We have applied methods from Natural Language Processing, especially Information Extraction technology, to both monolingual and Multilingual Authoring.

Machine Learning Approaches to Human Dialogue Modelling

We describe two major dialogue system segments: the first is an analysis module that learns to assign dialogue acts from corpora, but on the basis of limited quantities of data, and up to what seems to be some kind of limit on this task, a fact we also discuss. Secondly, we describe a Dialogue Manager which uses a representation of stereotypical dialogue patterns that we call Dialogue Action Frames, which are processed using simple and well understood algorithms, which are adapted from their original role in syntactic analysis role, and which, we believe, generate strong and novel constraints on later access to incomplete dialogue topics.

Knowledge-based multilingual document analysis

The growing availability of multilingual resources, like EuroWordnet, has recently inspired the development of large scale linguistic technologies, e.g. multilingual IE and Q&A, that were considered infeasible until a few years ago. In this paper a system for categorisation and automatic authoring of news streams in different languages is presented. In our system, a knowledge-based approach to Information Extraction is adopted as a support for hyperlinking. Authoring across documents in different languages is triggered by Named Entities and event recognition. The matching of events in texts is carried out by discourse processing driven by a large scale world model. This kind of multilingual analysis relies on a lexical knowledge base of nouns(i.e. the EuroWordnet Base Concepts) shared among English, Spanish and Italian lexicons. The impact of the design choices on the language independence and the possibilities it opens for automatic learning of the event hierarchy will be discussed.