Consensus reasoning for intelligent web searches
Suppose you wanted to know the answer to a specific question, such as "What is the atomic mass of Copper?" or "What is the diameter of Pluto?" Certainly, you could Google it and get many different answers. You could also ask experts or go to specific websites that you would expect to know the answer, such as NASA or Wikipedia, or you could even run a piece of software to get the answer. Now, suppose you were able to ask all of these sources at the same time. You'd probably find that you'd get many different, often conflicting, answers. How can those answers be combined into a single "best" consensus answer that is most likely to be closest to the "true" answer? With the immense amount of information available on the internet, it is possible to ask a single question and receive many different answers immediately. Expecting any single source to have the best answers for a wide range of questions or expecting all the sources to agree is quite unreasonable. The ideal approach is to collect answers from multiple sources, each with differing degrees of reliability, and consolidate them together into a single answer. The fundamental thesis of this research is that answer accuracy can be improved by processing candidate answers from multiple sources, using suitable methods to extract answers, detect consensus, handle conflicts, and identify source reliability.