Ghl George Fletcher

On the Expressive Power of the Relational Algebra on Finite Sets of Relation Pairs

We give a language-independent characterization of the expressive power of the relational algebra on finite sets of source-target relation instance pairs. The associated decision problem is shown to be co-graph-isomorphism hard and in co NP. The main result is also applied in providing a new characterization of the generic relational queries.

A DSL based on CSS for hypertext adaptation

Personalization offered by Adaptive Hypermedia and Recommender Systems is effective for tackling the information overload problem. However, the development of Adaptive Web-Based Systems is cumbersome. In order to ease the development of such systems, we propose a language based on CSS to express personalization in web systems that captures current adaptation techniques.

Efficient processing of containment queries on nested sets

We study the problem of computing containment queries on sets which can have both atomic and set-valued objects as elements, i.e., nested sets. Containment is a fundamental query pattern with many basic applications. Our study of nested set containment is motivated by the ubiquity of nested data in practice, e.g., in XML and JSON data management, in business and scientific workflow management, and in web analytics. Furthermore, there are to our knowledge no known efficient solutions to computing containment queries on massive collections of nested sets. Our specific contributions in this paper are: (1) we introduce two novel algorithms for efficient evaluation of containment queries on massive collections of nested sets; (2) we study caching and filtering mechanisms to accelerate query processing in the algorithms; (3) we develop extensions to the algorithms to a) compute several related query types and b) accommodate natural variations of the semantics of containment; and, (4) we present analytic and empirical analyses which demonstrate that both algorithms are efficient and scalable.

Stability notions in synthetic graph generation: a preliminary study

With the rise in adoption of massive graph data, it be- comes increasingly important to design graph processing algorithms which have predictable behavior as the graph scales. This work presents an initial study of stability in the context of a schema-driven synthetic graph generation. Specifically, we study the design of algorithms which gener- ate high-quality sequences of graph instances. Some desir- able features of these sequences include monotonic contain- ment of graph instances as they grow in size and consistency of structural properties across the sequence. Such stabil- ity features are important in understanding and explaining the scalability of many graph algorithms which have cross- instance dependencies (e.g., solutions for role detection in dynamic networks and graph query processing). We imple- ment a preliminary approach in the recently proposed open- source synthetic graph generator gMark and demonstrate its viability in generating stable sequences of graphs.

On structure preserving sampling and approximate partitioning of graphs

Massive graphs are becoming increasingly common in a variety of domains such as social networks and web analytics. One approach to overcoming the challenges of size is to sample the graph, and perform analytics on the smaller graph. However, to be useful, the sample must maintain the properties of interest in the original graph. In this paper, we analyze the quality of five representative sampling algorithms in how well they preserve graph structure, the bisimulation structure of graphs in particular. As part of this study, we also develop a new scalable algorithm for computing bisimulation partitions of massive graphs. We empirically demonstrate the superior performance of our new algorithm in both sequential and distributed settings.

Efficient RDFS entailment in external memory

The entailment of an RDF graph under the RDF Schema standard can easily become too costly to compute and maintain. It is often more desirable to compute on-demand whether a triple exists in the entailment. This is a non-trivial task likely to incur I/O costs, since RDF graphs are often too large to fit in internal memory. As disk I/O is expensive in terms of time, I/O costs should be minimized to achieve better performance. We investigate three physical indexing methods for RDF storage on disk, comparing them using the state of the art RDF Schema entailment algorithm of Munoz et al. In particular, the I/O behavior during entailment checking over these graph representations is studied. Extensive empirical analysis shows that an enhanced version of the state of the art indexing method, which we propose here, yields in general the best I/O performance.