The field of archaeology has long dealt with complex datasets, whether in the collection and analysis of archaeological finds, inter-site comparisons, regional analysis, evolutionary change, or in the framing of research questions which depend on multiple variables, factors, or other sparse and noisy datasets. The popularization of formal methods in the field has resulted in a variety of methods toward addressing these questions, largely borrowed from other fields, which should be subjected to a critical evaluation and discussion. To give just one example, the comparison of archaeological assemblages has been undertaken using a host of mathematical approaches: clustering techniques, principal components analysis, correspondence analysis, similarity coefficients, evolutionary models, and network analysis, and thus raises questions about the relationships between these different methods and their application to the same problem. The problem of inference, too, lurks in the background, and finding ways to evaluate a measure of confidence, certainty, or credibility in the results of modeling complex systems is in need of discussion. These issues can become ramified when dealing with methods which are predicated on higher numbers of variables. The aim of this workshop is therefore to bring awareness to the variety of techniques, engage in critical comparison and evaluation of the ways in which these tools are used, and examine how we should treat model selection, evaluation, and statistical inference in complex problems in archaeology.