The paper proposes a sentence alignment method based on maximum entropy model using anchor sentences to align ancient and modern Chinese sentences in historical classics. The method selects the sentence pairs with the same phrases at the beginning or the end of the sentence or with the same time phrases as anchor sentence pairs, which are employed to divide the paragraph into several sections. ...

This paper describes a hybrid text summarization method based on a TF-based sentence extraction method and a LEAD sentence extraction method. The LEAD method is known to be effective than other methods for document summarization of newspapers in lower summarization (output-to-input) ratio. In order to combine the LEAD method with the TF method, we used a rectangular distribution function that d...

We explore using relevant tweets of a given news article to help sentence compression for generating compressive news highlights. We extend an unsupervised dependency-tree based sentence compression approach by incorporating tweet information to weight the tree edge in terms of informativeness and syntactic importance. The experimental results on a public corpus that contains both news articles...

The issue about sentence similarity is essential to many areas of artificial intelligence. Although there are related studies on determining text similarity, fewer publications are about the similarity between short texts especially about the similarity between sentence pairs. In this paper, a novel method is proposed to estimate the sentence similarity with consideration of direct relevance an...

This paper proposes a new automatic speech summarization method having two stages: important sentence extraction and sentence compaction. Relatively important sentences are extracted from the results of large-vocabulary continuous speech recognition (LVCSR) based on the amount of information and the confidence measures of constituent words. The set of extracted sentences is compressed by our se...

Conventional sentence compression methods employ a syntactic parser to compress a sentence without changing its meaning. However, the reference compressions made by humans do not always retain the syntactic structures of the original sentences. Moreover, for the goal of ondemand sentence compression, the time spent in the parsing stage is not negligible. As an alternative to syntactic parsing, ...

The approaches previously used for sentence alignment (sentence length, word correspondence and cognate matching) take into account different aspects of similarity between the source and the target language sentences. In this paper we discuss various aspects of similarity in translated texts that can be used for sentence alignment. We then describe a customizable method for combining several ap...

This paper proposes a new automatic speech summarization method having two stages: important sentence extraction and sentence compaction. Relatively important sentences are extracted based on the amount of information and the confidence measures of constituent words, and the set of extracted sentences is compressed by our sentence compaction method. The sentence compaction is performed by selec...

∗ In Edmundson (1969) four features are used: the title feature, the cue word feature, the location feature and the word frequency feature. The frequency method Edmundson applied used the frequency of relevant words (frequency larger than a certain threshold and not being a common word) and assigned a score to each sentence based on the frequency of the relevant words in the sentence. Because o...