Aims & Topics of the Workshop

Over the last decades, data applications and, in particular, technological developments in scientific and Internet domains have generated very large data sets. Such data are often modeled as matrices. For example, term-document matrices can be constructed in many Internet applications indicating the frequency of one term in one document.

Such data matrices often have structural properties that present challenges and opportunities for researchers in Numerical Linear Algebra and Computer Science.

In fact, problems of information retrieval and Web analysis, such as Google PageRank problem and latent semantic indexing, have strongly motivated the research in the field of design and analysis of linear algebra algorithms involving massive data sets.

The workshop is aimed to explore novel techniques for modeling and analyzing massive, high-dimensional, and nonlinear-structured data, and to discuss new efficient and reliable numerical linear algebra algorithms for Web and large scale applications, bringing together computer scientists, computational and applied mathematicians, statisticians, and practitioners to promote cross-fertilization of ideas.

Areas and topics of interest for the workshop include (but are not limited to):

• Information retrieval on the Web
• Web search ranking methods
• Parameter selection in Web search-type matrices
• Sensitivity of models with respect to parameters and their interpretation
• Model hierarchies, model assembly and representation
• Iterative solution of Markov chains, including aggregation
• techniques, multigrid and algebraic domain decomposition
• Numerical methods for the singular value decomposition
• Numerical methods for large sparse or structured linear systems