Research student profile: Jason Morris
Generalized method of inferring data via rule-based pattern matching over pedotransfer functions
My research concerns the generalization of the methods and apparatus used in implementing the Soil Inferencing System (SINFERS) to other domains. The hypothesis is that machine learning techniques coupled with modern rule engines can be used to automate predictions from sparse data sets at considerable savings in time, cost, and effort. Predicted data can then be used in downstream decision support and planning systems. Such an approach has wide applicability to many sub-domains in the natural sciences and possibly other domains.
SINFERS is continuation of work done between G. Tranter and J. Morris (2009), in collaboration with CSIRO. It will be used to predict soil properties as input to future soil assessment systems. The system is implemented primarily in Java and deployed as a web application which any user can access with a web browser. SINFERS works by applying rules to pedotransfer functions (PTFs): selecting which PTFs to use, resolving rule conflicts, and checking for runtime errors. The project is an interesting example of how mature AI techniques can be applied to specific domains to produce useful tools.
I was born and raised in Michigan, USA. A mechanical engineer by schooling and a software engineer by trade, I’ve worked in IT and software engineering since 1997. Java technologies have been my mainstay since 2000, and rule-based expert systems my specialty since 2003. Specifically, I’m interested in applying rule-based technologies to scientific and academic computing as well as to web applications. I also like studying other applied AI subjects like machine learning, neural nets, and case-based reasoning. Currently, I hold the F. H. Loxton Scholarship in the Faculty of Agriculture and Environment in the Department of Environmental Sciences.
Outside of research, I do private IT consulting and serve as chief organizer of the IntelliFest: International Conference on Reasoning Technologies. View my full professional profile at LinkedIn.
In our department, I helped start a Python user group for our soil science grad students. When I’m not coding, I can be found hiking in the bush (here or in the States), playing chess, painting watercolors or fishing.
- The Role of Soil Inference Systems in Digital Soil Assessments, Digital Soil Assessments and Beyond: Proceedings of the 5th Global Workshop on Digital Soil Mapping 2012, Sydney, Australia