Automatic Generation of Text-Based Open Cloze Exercises
This article presents an approach to the automatic generation of open cloze exercises that are based on real-life English texts. The exercise format is similar to the open cloze test used in Cambridge certificate exams (FCE, CAE, CPE). Two experiments were conducted to evaluate the usefulness on the machine-generated exercises and compare them with authentic Cambridge tests. The experiments showed that the generation method used was quite effective. With some customization, the presented method can be applied to generating similar exercises based on texts written in other languages.
This article presents an approach to the automatic generation of open cloze exercises based on arbitrary English text. The exercise format is similar to the open cloze test used in Cambridge English certificate exams (FCE, CAE, CPE). The presented method also makes it possible to adjust the difficulty of the resulting exercises to better suit specific proficiency levels. Three experiments were conducted to evaluate the usefulness of the machine-generated exercises, compare them with authentic Cambridge English tests and study the difficulty-setting capabilities. The experiments showed that the generation method used was quite effective. With some customization, the method can be applied to generating similar exercises for other languages.
Academic environment constraints, i.e. limited amount of EFL hours for non-majors, call for a careful consideration of the stakeholders' needs in curriculum design in higher education context of a non-English speakign country. Tailoring an EFL course to those needs lands itself into an integrated program, comprising several study areas. This article discusses a case of syllabus design, combining a traditional EFL component, academic skills training and major-specific content (Country Studies for International Affairs majors). The author discusses the application of a solf systems methodology (SSM) in the academic environment.
Mastering English through Global Debate brings together rhetorical traditions and the best practices of ESL instruction to facilitate Superior-level proficiency in the English language. Each chapter addresses a rich topic of debate, providing students with a set of prereading activities, texts covering both sides of a debate topic, and postreading comprehension and lexical development exercises - all of which foster the language and critical thinking skills needed for successful debates. A rhetorical methods selection in each chapter integrates language and practice and prepares students for end-of-chapter debates. Using debate to develop advanced proficiency in a second language is a method that is finding increased interest among instructors and students alike, in both synchronous online teaching and the individual classroom. Students are prepared to participate fully in debates with their classmates - at home, abroad, or both.
Language exercises are widely used in teaching foreign languages; yet, manually creating exercises is labor-intensive and time-consuming. This paper describes a method for automatically generating EFL wordbank cloze exercises. These are generated from arbitrary passages in English, which is an important advantage in terms of learner motivation; indeed, the content of the exercises can be tailored to learners’ interests. Another feature of the method is exercise difficulty adjustment. Unlike other systems, our algorithm does not rely on many external linguistic resources and can be thus more easily adapted to other languages. Two experiments were conducted to evaluate the proposed method. The experiments showed that our algorithm performs significantly better than the ‘naïve’ random-sample baseline and that its precision of making gaps is 97%.
The workshop series on NLP for Computer-Assisted Language Learning (NLP4CALL) is a meeting place for researchers working on the integration of Natural Language Processing and Speech technologies in CALL systems and exploring the theoretical and methodological issues arising in this connection. The papers in the proceedings volume from the third NLP4CALL workshop cover three main topic areas: resources for development of ICALL applications (e.g., learner corpora and coursebook corpora), tools and algorithms for the analysis of learner language (e.g., focusing on collocations, reading tasks, cloze items, pronunciation, spelling, level classification of learner production), and the generation of learning materials (e.g., exercise generators).
The paper discusses some educational features of a modern media-sharing tool (VoiceThread) and highlights its pedagogical affordances in foreign language education.
The paper describes the learner corpus composed of English essays written by native Russian speakers. REALEC (Russian Error-Annotated Learner English Corpus) is an error-annotated, available online corpus, now containing more than 200 thousand word tokens in almost 800 essays. It is one of the first Russian ESL corpora, dynamically developing and striving to improve both in size and in features offered to users. We describe our perspective on the corpus, data sources and tools used in compiling it. Elaborate self-made classification of learners’ errors types is thoroughly described. The paper also presents a pilot experiment on creating test sets for particular learners’ problems using corpus data.