COLLEAP - COntextual Language LEArning Pipeline

摘要

In this paper we present a concept as well as a prototype of a tool pipeline to utilize the abundant information available on the World Wide Web for contextual, user driven creation and display of language learning material. The approach is to capture Wikipedia articles of the user's choice by crawling, to analyze the linguistic aspects of the text via natural language processing and to compile the gathered information into a visually appealing presentation of enriched language information. The tool is designed to address the Japanese language, with a focus on kanji, the pictographic characters used in Japanese scripture.