Output details
11 - Computer Science and Informatics
University of Leicester
Abstraction and Training of Stochastic Graph Transformation Systems
<10>Aspects that could lead to future applications (1): the method of parameter fitting (to reproduce experimental observations) using Bayesian networks is useful more widely and could significantly reduce the effort of months of simulation and manual adjustment (2): the approach to abstraction of stochastic rule-based systems is transferrable to other formalism, such as Rewriting Logic (3): one could use similar methods for extracting quantitative models in SE from measurements on running software. Following the presentation and publication of the paper these aspects have elicited the interest of experts in stochastic modelling such as Talcott (SRI), Petre (Turku) and Czarnecki (Waterloo).