Archives

Please note that this newsitem has been archived, and may contain outdated information or links.

Best Paper Award at CoNLL for Sarah Hiller and Raquel Fernandez

Sarah Hiller and Raquel Fernández have received the Best Paper Award at this year's Conference on Computational Natural Language Learning (CoNLL-2016) for their work on corrective feedback in first language acquisition.

CoNLL, which was held in Berlin this August, is a top-tier conference in computational linguistics, at the interface of machine learning and psycholinguistics. This year, CoNLL received 186 submissions. The Best Paper Award was sponsored by Google.

The paper, which is a follow-up of Sarah Hiller's Master of Logic thesis, is the first large scale data-driven investigation of implicit corrections in child-adult dialogue. These are responses that reformulate an erroneous child utterance, thereby providing a correct counterpart, as in the following example:

  CHILD: I climb up daddy.
  FATHER: You did climb over daddy.

The study shows that such implicit corrective feedback is associated with a reduction of errors after a time lag of at least 7 months and thus has an impact on language learning.

The paper and the data are available at https://staff.science.uva.nl/r.fernandezrovira/cf.php

Please note that this newsitem has been archived, and may contain outdated information or links.