Paper: Gathering and Generating Paraphrases from Twitter with Application to Normalization

ACL ID W13-2515
Title Gathering and Generating Paraphrases from Twitter with Application to Normalization
Venue Building and Using Comparable Corpora
Session
Year 2013
Authors

We present a new and unique para- phrase resource, which contains meaning- preserving transformations between infor- mal user-generated text. Sentential para- phrases are extracted from a compara- ble corpus of temporally and topically related messages on Twitter which of- ten express semantically identical infor- mation through distinct surface forms. We demonstrate the utility of this new re- source on the task of paraphrasing and normalizing noisy text, showing improve- ment over several state-of-the-art para- phrase and normalization systems 1.