PANDA: A platform for academic knowledge discovery and acquisition

Abstract

Scientific literatures contain some academic knowledge which is interesting or valuable but previously unknown. For instance, an algorithm A proposed in one article might have association with algorithm B in another article, while algorithm B is designed based on the definition of C in a third article. Thus we can deduce the relationship A-C based on A-B and B-C. There are also other kinds of academic knowledge such as association between two research communities, historical evolvement of a research topics, etc. But with the exponential growth of research articles that usually published in Portable Document Format (PDF), to discover and acquire potential knowledge poses many practical challenges. Existing algorithmic methods can hardly extend to handle diverse journals and layouts, nor scale up to process massive documents. As crowdsourcing has become a powerful paradigm for problem-solving especially for tasks that are difficult for computer to resolve solely, we state the problem of academic knowledge discovery and acquisition using an hybrid framework, integrating the accuracy of human workers and the speed of automatic algorithms. We briefly introduce a Platform for Academic kNowledge Discovery and Acquisition (PANDA), our current system implementation, as well as some preliminary achievements and promising future directions.