Pragmatic web-scale reasoning
Jacopo Urbani (Vrije Universiteit Amsterdam)
Info
|
Date |
31st January 2012 (week 3, Hilary Term 2012) |
|
Time |
11:30 |
|
Place |
147 |
Abstract
The large amount of Semantic Web data and its fast growth pose a significant computational challenge in performing efficient and scalable reasoning. On a large scale, the resources of single machines are no longer sufficient and we are required to distribute the process to improve performance. This domain is an excellent use-case for the high-performance computing community because it presents many interesting challenges that can be solved with various parallel programming models.
In my talk, I will look at this problem from a very pragmatic perspective. I will first illustrate a few of these challenges and then present some work that we have been doing on this topic in the last few years. More in detail, I will describe a forward-chaining reasoner and a SPARQL engine based on the MapReduce programming model. These techniques have excellent scalability since they allow OWL reasoning and SPARQL querying over billions of triples. Also, I will talk about an hybrid-chaining framework based on Ibis, which is high-performance framework developed at VUA. I will conclude by presenting some experiments that we performed on real-world and benchmark data that illustrate strong and weak points of our approaches.
Further info
|
Related series |
|
