An Automatic Feedback System for Academic Writing

Writing is an important activity in academic and corporate environments. Feedback on the writing is key to improving the quality of documents, and for novice writers to develop their skills. This paper describes the architecture of a web-based framework for providing automated feedback. The framework allows writers to use different writing tools such as Wiki’s, cloud-based Google Docs or Word. The documents are then processed to produce a wide range of types of feedback spanning collaborative or individual writing activities, visual and text modalities, feedback can be on surface or content features, on the writing product (the final document) or the process. This range of possibilities requires an architecture that is highly flexible and extensible. The paper also provides a taxonomy of automated feedback systems that shows how other systems approached different aspects of writing support. The evaluation provides evidence on the extensibility of the framework and how it is used on realistic teaching scenarios. This is an internal report released for those using glosser. The system can be downloaded from http://www.glosserproject.org