Implementing an XML based Unified Knowledge Manipulation Languages for SOAR cognitive architecture

SOAR is a major exemplar of architectural approach to machine cognition with numerous applications including speech recognition, machine perception, robotic, and strategy planning. We have replaced language used in SOAR architecture with our own developed XML based unified knowledge manipulation language (UKML) to demonstrate this language as a shared platform for procedural knowledge representation in cognitive architectures, and enhance SOAR with some valuable new features including the ability of manipulating XML formatted factual knowledge, greatly improved code readability and organization, and elimination of some ambiguities of Soar traditional language. This paper is a report on this implementation of UMKL, enriched with some examples and their results.