1. What is AutoQuiz?
AutoQuiz is a personalized automatic quiz generation system that evaluates reading comprehension and assesses English proficiency of English learners. This system is capable of generating multiple-choice questions, inclusive of vocabulary, grammar and reading comprehension, with various levels of difficulty.
Nowadays in a ubiquitous English learning environment, self-motivated learners can learn English while they surf the Internet, reading English news or documents of their professional fields. By taking advantage of an auto-quiz system, learners can self-assess if they comprehend the meaning of any text they browse or read. In the same time, learners can as well absorb new vocabulary.
2. What have we done so far?
So far, we’ve developed and completed the automatic generation of three major types of multiple-choice questions: reading comprehension (mostly on adjective understanding), vocabulary assessment and grammar assessment. What’s more, we’ve proposed a personalized adaptive English learning scheme which assesses students’ English proficiency and controls the level of questions generated and assigning to the students.
Reading Comprehension (Adjective Understanding)
In the first part, at which our system is most adept, in order to discourage learners from answering questions by recalling memorized answers from the article they read, and to expose learners with more vocabularies, the answer of the final question is a substitute of the adjective being examined. The candidates of a substitute are gathered from WordNet and filtered by web corpus searching.
Based on the proposed methodology, the choice candidates that are not selected as the answer are still useful. They are more distracting than the dissimilar adjectives and can be used as distractors to improve the distinguishability of a question.
Given an article with electronic format, our system will parse each sentence with the help of Link Grammar Parser and the semantic network is then built according to the parsing result. WordNet and Google are consulted in order to generate answers and distractors of multiple-choice questions. The vocabulary appearing as an answer to the question is selected from the official English vocabulary list released by College Entrance Examination Center and the vocabulary level corresponding to that of the student. With such mechanism, students get to learn vocabulary of appropriate level.
On the other hand, the grammar part is generated using rule-based approach and pattern-matching method.
Personalized Adaptive English Learning
Our system provides automatic question generation on the basis of personalized and adaptive English learning, which is a critical part of our research. With a pre-test that assesses a student’s English proficiency, the student is categorized into one of the 1-6 levels.
According to the level of a student, our system is able to extract English news of the corresponding level and then generates questions based on the news as well as the student’s level. That is, questions generated by our system vary with the level of a student since it’s trivial that an overly simple or difficult question does little help.
While we continue our research interests mentioned above, currently, we’re also working on question generation on reading comprehension with the hope of generating context-related multiple-choice questions that help students grasp the main idea of an article or news.
3. Any activities?
In recent years, AutoQuiz has been providing automatic question generation on English news during every summer and winter vacation. This activity aims at encouraging senior high school students to read English news online as a habit and our system serves as a way to assess students’ reading comprehension.
Fortunately, with the promotion of English e-learning platform- IWiLL and the encouragement of some enthusiastic English teachers, during every period of news reading activity, averagely hundreds of students participate in reading English news regularly and fulfill a required amount of quizzes generated automatically by our system.
To our delight, such activity not only helps students in learning English but also provides valuable feedback that allows us to enhance our research quality. We’re looking forward to refining our current question generation system and broadening the variety of question types while we continue our efforts in the future research.