Research Expenditure Allocation (REA)

REA (Research Expenditure Allocation) is the name of a software currently being developed at the Icelandic Institute for Intelligent Agents (IIIM).

Summary: In this project, I design and develop of a software architecture (GitHub Link here), leveraging Python framework for 1) automatically extracting symbolic-numeric data from business data and 2) compensation assessment for business processes via a learning-based algorithm. The system is being developed entirely from the ground up.

Details:

The project starts with exploring existing and novel method for achieving algorithmic transparency in automated systems used for computing derived data in business processes. Algorithmic transparency is important in businesses, but has turned out to be difficult to achieve with contemporary artificial intelligence (AI) methods. In this project we evaluate methodologies for identifying and interpreting processing symbolic-numeric data from databases with minimal human intervention, while ensuring alignment with business process requirements. In the project, I have:

• Learned about algorithms for identifying symbolic-numeric data in databases suitable for computing derived business information.
• Constructed an outline for software architectures that can automatically extract symbolic-numeric data from business data, with minimal human intervention, and compute derived information necessary for a company’s operations.
• Evaluated the transparency of said software architectures through its implementation and summarized the principles that are most successful for ensuring their transparency, in light of an organization’s operational objectives.

GitHub Link: https://github.com/IIIM-IS/REA