I have just completed the requirements for a Masters degree in Computer Science, under the supervision of Prof. Ido Dagan and Dr. Meni Adler in the NLP Lab at the Department of Computer Science, Bar-Ilan University.
My research field is Natural Language Processing. My main research interest is Textual Entailment and my focus is on utilizing human judgement in order to evaluate and improve entailment rule-base quality.
Thesis title: Evaluating and Improving Inference-rules via Crowdsourcing.
Abstract: The potential contribution of inference rules to many semantic applications has motivated many works for their automatic acquisition. Unfortunately, the utilization of inference rule resources is currently somewhat limited, partly due to the considerable number of rules which are in fact invalid. In my thesis I propose a replicable and reliable framework for evaluating inference-rules using crowdsourcing. I then demonstrate how, using this framework, existing inference-rule resources can be evaluated and improved by computing an estimate of the accuracy of their rules. Part of this work was presented as a talk at ACL 2012 (the main NLP international conference).
Castle Cottage Chilham B&B