We propose a supervised, two-phase framework to address the problem of paraphrase recognition (PR). Unlike most PR systems that focus on sentence similarity, our framework detects dissimilarities between sentences and makes its paraphrase judgment based on the significance of such dissimilarities. The ability to differentiate significant dissimilarities not only reveals what makes two sentences...

The design and implementation of a paraphrase component for a natural language question-answering system (CO-OP) is presented. The component is used to produce a paraphrase of a user's question to the system, which is presented to the user before the question is evaluated and answered. A major point made is the role of given and new information in formulating a paraphrase that differs in a mean...

Lattice decoding in statistical machine translation (SMT) is useful in speech translation and in the translation of German because it can handle input ambiguities such as speech recognition ambiguities and German word segmentation ambiguities. We show that lattice decoding is also useful for handling input variations. Given an input sentence, we build a lattice which represents paraphrases of t...

Paraphrase extraction relying on a single factor such as distribution similarity or translation similarity might lead to the loss of some linguistic properties. In this paper, we propose a paraphrase extraction framework, which accommodates various linguistically motivated factors to optimize the quality of paraphrase extraction. The major contributions of this study lie in the augmentable para...

We present an approach for automatically learning synonyms from a corpus of paraphrased tweets. The synonyms are learned by using shallow parse chunks to create candidate synonyms and their context windows, and the synonyms are substituted back into a paraphrase detection system that uses machine translation metrics as features for a classifier. We find a 2.29% improvement in F1 when we train a...

We analyze in this paper a number of data sets proposed over the last decade or so for the task of paraphrase identification. The goal of the analysis is to identify the advantages as well as shortcomings of the previously proposed data sets. Based on the analysis, we then make recommendations about how to improve the process of creating and using such data sets for evaluating in the future app...

Abstract The Turku Paraphrase Corpus is a dataset of over 100,000 Finnish paraphrase pairs. During the corpus creation, we strived to gather challenging pairs, more suitable test capabilities natural language understanding models. paraphrases are both selected and classified manually, so as minimise lexical overlap, provide examples that structurally lexically different maximum extent. An impor...

We present a new release of the Paraphrase Database. PPDB 2.0 includes a discriminatively re-ranked set of paraphrases that achieve a higher correlation with human judgments than PPDB 1.0’s heuristic rankings. Each paraphrase pair in the database now also includes finegrained entailment relations, word embedding similarities, and style annotations.