The PEO project investigates methods for representation and interpretation of scientific and natural processes and events. Mammoth amounts of data, that pertains to solar activity, atmospheric conditions, ocean currents etc, is being collected at various observatories, and this project is one attempt to digest it. These observations are temporally and spatially related by the processes and events happening when the data was being collected. We want to infer these proce>sses and events by munching data using their machine readable descriptions.
The first goal is to develop a theory to interpret processes. The theory must help 1. decide whether a process can happen given a set of conditions, and if it can happen, 2. what is the effect that it will have, 3. being aware of other processes that may be occurring concurrently. Given current state and conditions the theory should help predict changes in the state if the process occurred. In contrast given some observations, the theory must help predict if the process could have played a role in changes noticed in the observations.
Next step would be to develop a language for representing processes such that the process descriptions can be unambiguously interpretted using the theory developed in the first step.
In the last step we would focus on the development of an efficient, scalable, open-source reasoner that can munch process descriptions in the language developed in the second phase of the project and observations to guess the processes and events that were in play when the data was collected, along with justifications for the drawn inferences.