Generating Comprehension Questions Using Paraphrase - Technologies and Applications of Artificial Intelligence - page 318

artificial statements that could test the knowledge of multiple spans of information. We
introduced the concept of paraphrase when designing the rules, allowing them to perform
paraphrasing actions. The Paraphrase Generation System includes paraphrase resources
that are suitable to our system. Particularly, we added QG-specific resource, nominal
coreference, to capture the article-wide coreferential relations. Finally, a two-way clas-
sifier, the Acceptability Ranker, was trained from an annotated data set generated by our
system. We integrated useful features from both rankers for question generation and
paraphrase generation. The experimental results suggest that our system are more capable
of generating challenging test choices that would not be simply solved by matching exact
word span and would be more likely to distinguish those who do not comprehend the
reading article well from those who do.
In the future, we plan to investigate the possibility of using implicit discourse rela-
tions and incorporate entailment-based rules into our system. We believe that implicit
discourse relations would test a higher level of comprehension than explicit ones be-
cause the former do not give obvious clues, like connectives. The idea of rewriting a
statement while pertaining/changing its correctness conforms to rewriting a statement
into another with/without an entailment relationship between them. Entailment is
expected to increase the variety of the generated statements. Ultimately, we hope to
develop directly applicable question generation system that benefits e-learning envi-
ronment in the near future.
Acknowledgement. This research was partially supported by National Science Coun-
cil of Taiwan under grant NSC100-2221-E-001-015-MY3.
References
1. Afzal, N., Mitkov, R.: Automatic generation of multiple choice questions using dependency-
based semantic relations. Soft Computing 18, 1269-1281 (2014)
2. Agarwal, M., Shah, R., Mannem, P.: Automatic question generation using discourse cues.
In: Proceedings of the 6th Workshop on Innovative Use of NLP for Building Educational
Applications, pp. 1-9. Association for Computational Linguistics (2011)
3. Bernhard, D., De Viron, L., Moriceau, V., Tannier, X.: Question Generation for French:
Collating Parsers and Paraphrasing Questions. Dialogue and Discourse 3(2), 43-74 (2012)
4. Bhagat, R., Hovy, E.: What is a paraphrase? Computational Linguistics 39(3), 463-472
(2013)
5. Ganitkevitch, J., Callison-Burch, C., Van Durme, B.: Ppdb: The paraphrase database. In:
Proceedings of NAACL-HLT, pp. 758-764 (2013)
6. Heilman, M.: Automatic Factual Question Generation from Text. Ph.D. Dissertation,
Carnegie Mellon University. CMU-LTI-11-004 (2011)
7. Heilman, M., Smith, N.A.: Extracting Simplified Statements for Factual Question
Generation. In: Proceedings of QG 2010: The Third Workshop on Question Generation,
pp. 11-20 (2010)
8. Heilman, M., Smith, N.A.: Good question! statistical ranking for question generation. In:
NAACL-HLT, pp. 609-617 (2010)