Automatic paraphrase acquisition from news articles

Abstract

Paraphrases play an important role in the variety and complexity of natural language documents. However they adds to the difficulty of natural language processing. Here we describe a procedure for obtaining paraphrases from news article. A set of paraphrases can be useful for various kinds of applications. Articles derived from different newspapers can contain paraphrases if they report the same event of the same day. We exploit this feature by using Named Entity recognition. Our basic approach is based on the assumption that Named Entities are preserved across paraphrases. We applied our method to articles of two domains and obtained notable examples. Although this is our initial attempt to automatically extracting paraphrases from a corpus, the results are promising.

Unsupervised Dependency Parsing with Transferring Distribution via Parallel Guidance and Entropy Regularization

“…In recent years, dependency parsing has gained universal interest due to its usefulness in a wide range of applications such as synonym generation (Shinyama et al, 2002), relation extraction (Nguyen et al, 2009) and machine translation (Katz-Brown et al, 2011;Xie et al, 2011). Several supervised dependency parsing algorithms (Nivre and Scholz, 2004;McDonald et al, 2005a;McDonald et al, 2005b;McDonald and Pereira, 2006;Carreras, 2007;Koo and Collins, 2010;Ma and Zhao, 2012; have been proposed and achieved high parsing accuracies on several treebanks, due in large part to the availability of dependency treebanks in a number of languages .…”

Using Verbs to Characterize Noun-Noun Relations

“…For example, for "X solves Y" they extract paraphrasing templates like "Y is resolved by X", "X resolves Y", "X finds a solution to Y" and "X tries to solve Y". The idea is extended by Shinyama et al [22], who use named entities of matching semantic class as anchors, e.g., LOCATION, ORGANIZATION, etc. However, the goal of these approaches is to create summarizing paraphrases, while we are interested in finding noun compound semantics.…”

“…We also omitted charitable donation and Sem. relation Example Extracted Verbs POSSESSION family estate be in (29), be held by (9), be owned by(7) TEMPORAL night flight arrive at (19), leave at (16), be at (6), be conducted at(6), occur at(5) IS-A(HYPERNYMY) Dallas city include(9) CAUSE malaria mosquito carry(23), spread (16), cause (12), transmit (9), bring (7), have(4), be infected with(3), be responsible for (3), test positive for(3), infect many with(3), be needed for(3), pass on(2), give(2), give out(2) MAKE/PRODUCE shoe factory produce(28), make(13), manufacture(11) INSTRUMENT pump drainage be controlled through(3), use(2) LOCATION/SPACE Texas university be(5), be in(4) PURPOSE migraine drug treat(11), be used for (9), prevent (7), work for(6), stop(4), help(4), work(4) be prescribed for(3), relieve(3), block(3), be effective for(3), be for (3), help ward off (3), seem effective against(3), end(3), reduce(2), cure(2) SOURCE olive oil come from (13), be obtained from (11), be extracted from(10), be made from (9), be produced from (7), be released from(4), taste like(4), be beaten from(3), be produced with(3), emerge from(3) TOPIC art museum focus on (29), display(16), bring (14), highlight(11), house(10), exhibit (9) demonstrate (8), feature (7), show (5), tell about(4), cover(4), concentrate in(4) MEANS bus service use (14), operate(6), include(6) EXPERIENCER disease victim spread(12), acquire (12), suffer from(8), die of (7), develop(7), contract(6), catch(6), be diagnosed with(6), have(5), beat(5), be infected by(4), survive(4), die from(4), get(4), pass(3), fall by(3), transmit(3), avoid(3) THEME car salesman sell(38), mean inside(13), buy (7), travel by(5), pay for(4), deliver(3), push(3), demonstrate(3), purr(3), bring used (3), know more about(3), pour through(3) RESULT combustion gas support (22), result from (14), be produced during(11), be produced by(8), be formed from (8), form during (8)...…”

Building a Non-Trivial Paraphrase Corpus Using Multiple Machine Translation Systems

“…For instance, paraphrases have been acquired from monolingual comparable corpora such as news articles regarding the same event (Shinyama et al, 2002) and multiple definitions of the same concept (Hashimoto et al, 2011). Although these methods effectively acquire paraphrases, there are not many domains that have comparable corpora.…”