Paraphrase generation is an important task that has received a great deal of interest recently. Proposed data-driven solutions to the problem have ranged from simple approaches that make minimal use of NLP tools to more complex approaches that rely on numerous language-dependent resources. Despite all of the attention, there have been very few direct empirical evaluations comparing the merits of the different approaches. This paper empirically examines the tradeoffs between simple and sophisticated paraphrase harvesting approaches to help shed light on their strengths and weaknesses. Our evaluation reveals that very simple approaches fare surprisingly well and have a number of distinct advantages, including strong precision, good coverage, and low redundancy. 1

Paraphrasing methods recognize, generate, or extract phrases, sentences, or longer natural language expressions that convey almost the same information. Textual entailment methods, on the other hand, recognize, generate, or extract pairs of natural language expressions, such that a human who reads (and trusts) the first element of a pair would most likely infer that the other element is also true. Paraphrasing can be seen as bidirectional textual entailment and methods from the two areas are often very similar. Both kinds of methods are useful, at least in principle, in a wide range of natural language processing applications, including question answering, summarization, text generation, and machine translation. In this thesis, we focus on paraphrase and textual entailment recognition, as well as paraphrase generation. We propose three paraphrase and textual entailment recognition methods, experimentally evaluated on existing benchmarks. The key idea is that by capturing similarities at various abstractions of the inputs, we can recognize paraphrases and textual entailment reasonably well. Additionally, we exploit WordNet and use features that operate on the syntactic level of the language expressions. The best

Sentence compression has attracted much interest in recent years, but most sentence compressors are extractive, i.e., they only delete words. There is a lack of appropriate datasets to train and evaluate abstractive sentence compressors, i.e., methods that apart from deleting words can also rephrase expressions. We present a new dataset that contains candidate extractive and abstractive compressions of source sentences. The candidate compressions are annotated with human judgements for grammaticality and meaning preservation. We discuss how the dataset was created, and how it can be used in generate-and-rank abstractive sentence compressors. We also report experimental results with a novel abstractive sentence compressor that uses the dataset. 1

We present a method that paraphrases a given sentence by first generating candidate paraphrases and then ranking (or classifying) them. The candidates are generated by applying existing paraphrasing rules extracted from parallel corpora. The ranking component considers not only the overall quality of the rules that produced each candidate, but also the extent to which they preserve grammaticality and meaning in the particular context of the input sentence, as well as the degree to which the candidate differs from the input. We experimented with both a Maximum Entropy classifier and an SVR ranker. Experimental results show that incorporating features from an existing paraphrase recognizer in the ranking component improves performance, and that our overall method compares well against a state of the art paraphrase generator, when paraphrasing rules apply to the input sentences. We also propose a new methodology to evaluate the ranking components of generate-and-rank paraphrase generators, which evaluates them across different combinations of weights for grammaticality, meaning preservation, and diversity. The paper is accompanied by a paraphrasing dataset we constructed for evaluations of this kind. 1

Document revision histories are a useful and abundant source of data for natural language processing, but selecting relevant data for the task at hand is not trivial. In this paper we introduce a scalable approach for automatically distinguishing between factual and fluency edits in document revision histories. The approach is based on supervised machine learning using language model probabilities, string similarity measured over different representations of user edits, comparison of part-of-speech tags and named entities, and a set of adaptive features extracted from large amounts of unlabeled user edits. Applied to contiguous edit segments, our method achieves statistically significant improvements over a simple yet effective edit-distance baseline. It reaches high classification accuracy (88%) and is shown to generalize to additional sets of unseen data. 1

Predicting entailment between two given texts is an important task upon which the performance of numerous NLP tasks depend on such as question answering, text summarization, and information extraction. The degree to which two texts are similar has been used extensively as a key feature in much previous work in predicting entailment. However, using similarity scores directly, without proper transformations, results in suboptimal performance. Given a set of lexical similarity measures, we propose a method that jointly learns both (a) a set of non-linear transformation functions for those similarity measures and, (b) the optimal non-linear combination of those transformation functions to predict textual entailment. Our method consistently outperforms numerous baselines, reporting a micro-averaged F-score of 46.48 on the RTE-7 benchmark dataset. The proposed method is ranked 2-nd among 33 entailment systems participated in RTE-7, demonstrating its competitiveness over numerous other entailment approaches. Although our method is statistically comparable to the current state-of-the-art, we require less external knowledge resources.