صفحات جديدة باللغة العربية حصريًا قريبًا

يسرّنا الإعلان أننا نعكف حاليًا على إعداد صفحات جديدة مُصمّمة لجمهورنا الناطق باللغة العربية لتقديم تجربة استخدام متميزة ومحتوى مخصص وملائم أكثر لهم.

سنطلق هذه الصفحات المرتقبة قريبًا في الأشهر القليلة

Dedicated Arabic Pages Are Coming Soon

We're excited to announce that we are actively developing new, dedicated pages specifically designed for our Arabic-speaking users. These will offer tailored content and an enhanced experience.

Expected to launch in the next few months. Stay tuned!

Testing English Reading Comprehension through Deep Text Analysis and Question Generation

Kemal Oflazer

CMU-Q Point of Contact

This is a follow-up proposal to NPRP-09-873-1-129 “A natural language processing-based active and interactive platform for accessing English language content and advanced language learning” (completed August 2014) whose goal was to develop a English reading tutoring system called SmartReader. The SmartReader system provides active and interactive reading environment for English text making it possible for a student to interact with a text in a variety of ways using anytime-anywhere contextually guided access to information, such as word and phrase level operations, resolving anaphora, sentence simplification, name tracking, question generation and answering, summarization etc. In this follow-up project proposal, we would like to build up the expertise gained in this project and other work that we have been doing in deep question generation. In order to generate more sophisticated level of questions to see if the students do understand the texts not only syntactically but also semantically, we need to develop semantically motivated question generation from the multiple sources of sentences, using coreference resolution and detailed analysis of text, such as event semantics, inferences, paraphrases and semantic roles (e.g. agent, patient, time, location). The goal is to develop algorithm and technologies for analyzing short fragments of texts consisting of a sequence of sentences and generate questions whose answers require bring together information from multiple sentences using a nontrivial reasoning process. One very general and useful application area for this technology is to develop a tool that students can use to generate questions at their own pace and to practice exams for English comprehension test that is very similar to questions found in exams like TOEFL, SAT Verbal or ACT. We believe this could be deployed in Qatari high schools to improve the quality of learning English as a second language. The general techniques developed in this project can in the future be applied to a similar application working on Arabic texts.

Project

NPRP 8-1337-1-243

Year

2016

Status

Closed

Team
image

Teruko Mitamura

Carnegie Mellon University