The Military Language Tutor (MILT) Program: An Advanced Authoring System
Jonathan D. Kaplan, Mark A. Sabol, Robert A. Wisher, Robert J. Seidel
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
The Military Language Tutor (MILT) is a language tutor authoring system developed by the U.S. Army Research Institute (ARI). MILT gives authors the ability to develop tutoring programs with various functional capabilities, including natural language processing (NLP), two-dimensional, animated, microworld simulation, performance-based, instructional branching, presentation of most conventional types of questions and feedback, etc. Currently it contains two relatively elaborate and usable examples of lessonware on which authors can build — one in Arabic and the other in Spanish. Based on the interest of the Special Operations forces (SOF), we developed a proof of principal version of MILT's two-dimensional Arabic microworld which uses speech rather than keyboard input to control an animated agent in solving an authored problem. The speech is recognized using a Dragon discrete speech recognizer utilizing a relatively small modern standard Arabic speech model. It is this speech-driven microworld that was the primary focus of an evaluation performed at Fort Campbell, KY. This version of the MILT microworld is controlled by 70 spoken Arabic sentences. The limitations of discrete speech recognition means that spontaneous speech recognition is impossible and the expected utterances must be presented for reading. As a result, three of the Arabic sentences are written at a time on the bottom of the microworld screen and students read aloud the sentence that they wish to see executed. When such a command is recognized and executed, the system automatically replaces the sentences with another set of three, and so forth. The evaluation produced significant, positive learning effects and useful attitudinal information from students. As a result, the discrete speech recognizer in MILT is being replaced by a continuous recognizer, and the two-dimensional microworld is being replaced by a three-dimensional one.
逐年被引趋势
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
计算机 / AISpeech and dialogue systems
Natural Language Processing Techniques · Intelligent Tutoring Systems and Adaptive Learning
参考文献 0
引用本文 21
按被引量排序,此处列出前 3 条