Microsoft Research has published a new paper entitled “Pedagogy: a social networking approach to knowledge discovery.” The summary states, “The construction of semantically structured, accurate, and comprehensive knowledge bases remains an important but difficult problem. Acquiring the wealth of human knowledge in a machine-readable format would have invaluable benefits for semantic web services and AI applications in general. In this report I outline a new approach that aims to address this challenge by harnessing the power of social networking.”The paper is written by Mohammad Raza of Microsoft Research Cambridge. It begins, “A number of projects have made significant progress toward the goal of knowledge discovery, and these may be broadly classed into two categories. One approach is to develop methods for extracting knowledge from latent sources that currently exist on the web in human-readable form, such as Wikipedia. Examples of such projects include Yago [14,15], DBPedia [13], True Knowledge [12] and Powersetʼs ‘quick facts’ [11]. The challenge in this case is to develop sophisticated extraction techniques that can both guarantee accuracy and also deliver comprehensive coverage. However, existing methods need to trade one for the other, and the problem seems to be AI complete in general.” Download the paper for free to read more.
