Generating Rules with Common Knowledge: A Framework for Sentence Information Extraction

TL;DR: This work proposes an approach to combine the common knowledge and the nature language processing rules that first applied the name entity reorganization technology and then generated rules based on a specific common knowledge database.

Abstract: There are many nature language processing applications. A typical example is information extraction whose target is a sentence. Various rules are often used in this kind of applications. However, automated processing is not accurate enough in some cases. This is because it is easy to construct syntax rules of a sentence but difficult to semantic rules. On the other hand, the knowledge representation community paid much attention to common knowledge. It is insightful to use rules based on this sense on common things in nature language processing. Therefore, we propose an approach to combine the common knowledge and the nature language processing rules. It first applied the name entity reorganization technology and then generated rules based on a specific common knowledge database. As a result, this approach can be a framework for many (but not all) nature language processing applications. In our experimental example, this approach performed well.