Knowledge-intensive question answering

With recent advances in computer and Internet technology, people have access to more information than ever before. Much of the information is available in free text with little or no metadata, and there is a tremendous need for tools to help organize, classify, and store the information, and to allow better access to the stored information. Research in information retrieval (IR) has made much progress in addressing this problem. However, current IR systems only allow us to locate documents that might contain the pertinent information; most of them leave it to the user to extract the useful information from a ranked list. This leaves the (often unwilling) user with a relatively large amount of text to consume. People have questions and they need answers, not documents. Corpus-based question answering is designed to take a step closer to information retrieval rather than document retrieval. Briefly, the question answering (QA) task is to find, in a large collection of data, an answer to a question posed in natural language. Here’s an example of a fact-based question that modern corpusbased QA systems are able to answer by returning a short text snippet (taken from a document in the collection) that is believed to contain an answer.