We use probabilistic reasoning methods and learning algorithms to understand data , discover patterns, correlations and causal relationships. We then develop tools and simulation models to identify and evaluate options to improve outcomes.
Our Product: Policymakers’ Workbench
The Policymakers’ Workbench, is a knowledge-based software engine that synthesizes a large body of research and experimental studies and uses a probabilistic reasoning methodology to estimate the impact and effectiveness of various policy options. The Policymakers’ Workbench consists of two main components: (i) a model; and (ii) a computer user interface.
The model represents knowledge in the field and is built upon a set of underlying relations or functions. The theoretical engine or model is built upon a set of underlying relations or functions, such as “if student attendance goes up, achievement goes up, but only if there is time on task.” This model is elaborated as a Bayesian network that specifies probabilistic dependencies among the variables.
The computer interface enables users to evaluate the effect of changes in specific variables; perform diagnostics; and analyze the impact of various interventions or policies. The organization of beliefs in networks makes it possible to update or improve the entire system of beliefs by introduction of new evidence about individual variables.