Cognitive Architectures
Mike Byrne

1. Introduction
   • What are "cognitive architectures"
     • Cognitive Science theories of human cognition
     • Instantiated as software programs
   • Why are they relevant to HCI
     • Vision of "virtual user"
     • Short of that vision, ability to make useful predictions about performance/learnability
2. Past use in HCI
   • Basis for GOMS and MHP
   • CCT and transfer of training
   • Other examples
     • PUMs
     • Analyses of ATMs (CHI'95 workshop)
     • Interface exploration (Boulder Cricket Graph stuff)
     • Others as space allows
3. Recent Architectures
   • LICAI
     • Kitajima & Polson
     • Based on Construction-Integration, cognition as comprehension
     • Applications
   • EPIC
     • Kieras & Meyer (and others, e.g. Hornof)
     • Extension and implementation of MHP
     • Applications
       • Multiple-task performance
       • Telephone operators
       • Others as space allows
   • ACT-R/PM
     • Byrne and Anderson (and others)
     • Union of ACT cognitive theory with EPIC-like peripherals
     • Applications
       • ATC-like applications
       • Incidental location learning (Ehret thesis)
4. Future of Cognitive Architectures in HCI
   • Higher fidelity
   • More complex tasks
     • More problem-solving
     • Group tasks (e.g. CSCW)
   • Linking to real systems
     • CMIMS
     • GLEAN
     • St. Amant's work
   • Other challenges
     • Difficult to learn/use
     • Limited by the underlying theories (e.g. no emotion)