We introduce a stochastic graph-based method for computing relative importance of textual units for Natural Language Processing. We test the technique on the problem of Text Summarization (TS). Extractive TS relies on the concept of sentence salience to identify the most important sentences in a document or set of documents. Salience is typically defined in terms of the presence of particular important words or in terms of similarity to a centroid pseudo-sentence. We consider a new approach, LexRank, for computing sentence importance based on the concept of eigenvector centrality in a graph representation of sentences. In this model, a connectivity matrix based on intra-sentence cosine similarity is used as the adjacency matrix of the graph representation of sentences. Our system, based on LexRank ranked in first place in more than one task in the recent DUC 2004 evaluation. In this paper we present a detailed analysis of our approach and apply it to a larger data set including data from earlier DUC evaluations. We discuss several methods to compute centrality using the similarity graph. The results show that degree-based methods (including LexRank) outperform both centroid-based methods and other systems participating in DUC in most of the cases. Furthermore, the LexRank with threshold method outperforms the other degree-based techniques including continuous LexRank. We also show that our approach is quite insensitive to the noise in the data that may result from an imperfect topical clustering of documents.
Quickly moving to a new area of research is painful for researchers due to the vast amount of scientific literature in each field of study. One possible way to overcome this problem is to summarize a scientific topic. In this paper, we propose a model of summarizing a single article, which can be further used to summarize an entire topic. Our model is based on analyzing others' viewpoint of the target article's contributions and the study of its citation summary network using a clustering approach.
In this paper, we describe a system to rank suspected answers to natural language questions. We process both corpus and query using a new technique, predictive annotation, which augments phrases in texts with labels anticipating their being targets of certain kinds of questions. Given a natural language question, an IR system returns a set of matching passages, which are then analyzed and ranked according to various criteria described in this paper. We provide an evaluation of the techniques based on results from the TREC Q&A evaluation in which our system participated.
We present a multi-document summarizer, called MEAD, which generates summaries using cluster centroids produced by a topic detection and tracking system. We also describe two new techniques, based on sentence utility and subsumption, which we have applied to the evaluation of both single and multiple document summaries. Finally, we describe two user studies that test our models of multi-document summarization.
This paper presents the results of a study on the semantic constraints imposed on lexical choice by certain contextual indicators. We show how such indicators are computed and how correlations between them and the choice of a noun phrase description of a named entity can be automatically established using supervised learning. Based on this correlation, we have developed a technique for automatic lexical choice of descriptions of entities in text generation. We discuss the underlying relationship between the pragmatics of choosing an appropriate description that serves a specific purpose in the automatically generated text and the semantics of the description itself. We present our work in the framework of the more general concept of reuse of linguistic structures that are automatically extracted from large corpora. We present a formal evaluation of our approach and we conclude with some thoughts on potential applications of our method.
This paper explores morpho-syntactic ambiguities for French to develop a strategy for part-of-speech disambiguation that a) reflects the complexity of French as an inflected language, b) optimizes the estimation of probabilities, c) allows the user flexibility in choosing a tagset. The problem in extracting lexical probabilities from a limited training corpus is that the statistical model may not necessarily represent the use of a particular word in a particular context. In a highly morphologically inflected language, this argument is particularly serious since a word can be tagged with a large number of parts of speech. Due to the lack of sufficient training data, we argue against estimating lexical probabilities to disambiguate parts of speech in unrestricted texts. Instead, we use the strength of contextual probabilities along with a feature we call ``genotype'', a set of tags associated with a word. Using this knowledge, we have built a part-of-speech tagger that combines linguistic and statistical approaches: contextual information is disambiguated by linguistic rules and n-gram probabilities on parts of speech only are estimated in order to disambiguate the remaining ambiguous tags.
This paper addresses issues in part of speech disambiguation using finite-state transducers and presents two main contributions to the field. One of them is the use of finite-state machines for part of speech tagging. Linguistic and statistical information is represented in terms of weights on transitions in weighted finite-state transducers. Another contribution is the successful combination of techniques -- linguistic and statistical -- for word disambiguation, compounded with the notion of word classes.
In this paper, we describe a method for automatic creation of a knowledge source for text generation using information extraction over the Internet. We present a prototype system called PROFILE which uses a client-server architecture to extract noun-phrase descriptions of entities such as people, places, and organizations. The system serves two purposes: as an information extraction tool, it allows users to search for textual descriptions of entities; as a utility to generate functional descriptions (FD), it is used in a functional-unification based generation system. We present an evaluation of the approach and its applications to natural language generation and summarization.