For more information contact Luis Rocha at rocha@lanl.gov
Last Modified: September 11, 2002
An essential technique to understand and predict the dynamics of natural systems is the construction of models. Scientific models extract information from data to infer relationships between empirical patterns and the processes that produce them. Though the mechanisms by which different models process information vary according to the model category and application domain, similar challenges confront all such activities.
This talk will compare alternative approaches to modelling complex biological processes. The focus will be on how each modelling approach processes information to establish a representation of the regularities therein. Examples will be drawn from a range of applications, including agent-based models of ecological community assembly, mutual information analyses of coevolving organisms, and application of information-based model selection to understand the genetic components of natural variation in rates of HIV-1 infection progression. Contributions and limitations of each approach will be discussed in reference to the state of the art of constructing and validating scientific models.