《Auto-Approximation of Graph》.pdf

AutoApproximation of Graph Computing Zechao Shang Jeffrey Xu Yu The Chinese University of Hong Kong The Chinese University of Hong Kong Hong Kong, China Hong Kong, China zcshang@.hk yu@.hk ABSTRACT the final answer in an automatic and systematic way without de- In the big data era, graph computing is one of the challenging issues signing new approximate algorithms, which are extremely difficult because there are numerous large graph datasets emerging from real not only for end-users but also for graph algorithm experts? The applications. A question is: do we need to know the final exact an- main idea behind the questions is more data will beat clever algo- swer for a large graph? When it is impossible to know the exact rithms [13]. In other words, it becomes more important to find out answer in a limited time, is it possible to approximate the final an- something meaningful quick from a large graph, and we should fo- swer in an automatic and systematic way without having to design- cus on finding a way of making full use of large graphs instead of ing new approximate algorithms? The main idea behind the ques- spending time on designing approximate algorithms. tion is: it is more important to find out something meaningful quick However, this is not easy. We cannot understand what a pro- from a large graph, and we should focus on finding a way of making gram is trying to compute, even with access to the source co