TRAINING SYSTEMS FOR PROCEDURAL COGNITIVE TASKS

RECENT ADVANCES IN THE COGNITIVE, INSTRUCTIONAL AND COMPUTERSCIENCES PROVIDE AN OPPORTUNITY FOR THE DEVELOPMENT OF A NEWGENERATION OF TRAINING SYSTEMS AT LOWER COST AND GREATER EFFECTIVENESS THAN PREVIOUSLY POSSIBLE. SPECIFICALLY, THE PROPOSED EFFORT BUILDS ON TWO UNIQUE OPPORTUNITIES: (A) OUR GENERAL PURPOSE RULETUTOR, THE CONTENT-INDEPENDENT SOFTWARE NECESSARY FOR CONSTRUCTING ANY PROCEDURAL COGNITIVEPERFORMANCE ID/TRAINING SYSTEM AND (B) SPECIFICATIONS FOR A GENERAL METHOD OF COGNITIVE TASK ANALYSIS, CALLED STRUCTURALANALYSIS (SA). SA TAKES AS INPUT SPECIFIC TASK SOLUTIONS GENERATED BY EXPERTS AND GENERATES PSYCHOLOGICALLY RELEVANT PROCEDURAL SPECIFICATIONS (FOR GIVEN CLASSES OF TASKS). THE PROPOSED RESEARCH IS DESIGNED TO FULLY AUTOMATE SA AND TO INTERGRATE IT WITH THE GENERAL PURPOSE RULETUTOR. MORE SPECIFICALLY, WE PROPOSE TO USE THE RULETUTOR AS A BASIS FORAN ENTIRE CLASS OF PERFORMANCE AIDS AND INDIVIDUALIZED TRAINING SYSTEMS. THE MAJOR EFFORT WILL GO INTO AUTOMATING THE PROCESS OF SA AND DEVELOPING A VERY HIGH LEVEL RULE-ORIENTED LANGUAGE WHICH WILL TAKE THE RESULTS OF SA (I.E., PSYCHOLOGICALLY RELEVANT RULES AND RULE COMPONENTS) AND AUTOMATICALLY GENERATE (LOWER-LEVEL LANGUAGE) SOFTWARE CODE (E.G. PASCAL OR EVEN BASIC) AND INTEGRATE IT INTO THE GENERAL PURPOSE RULETUTOR.