A human behavior inspired awareness system for document analysis

With the rapid growth of the Internet, it is natural that we want to handle and process the document not as pen and paper but as online information. To deal with such large amount of information, people often want to use a certain classifier to classify it. To train a classifier, teacher signals are needed. However, different applications have different features, such as spam mail filter, patent analysis and hot-topic retrieval system. It is difficult to get enough teacher signals for every application. In this study, we describe a brain inspired awareness system for document analysis. This system could generate different document classifiers without any prior information. Then the user could give instructions to the system. After retraining, the system could understand user's intention and know how to generate a good classification system. Since the learning methods are drawn inspiration from human learning behaviors, which would be very easy for human beings to understand.