Augmenting and Automating Corpus Enrichment

Abstract

An agent in pursuit of a task may work with a reference library containing documents associated with additional data that provide location-specific explanations about the content. Faced with a new document, an agent has to decide whether to include the new document in its reference library. Basing the decision on words, topics, or entities has shown not to lead to a balanced performance for varying documents. In this paper, we present an approach for automatically enriching new documents with data associated to documents in a reference library. Additionally, we analyze these data to classify new documents into categories to help an agent in deciding whether to include the new document in its reference library.