Knowledge Representation

• In the Knowledge Representation lecture we talked about a range of techniques including:
  • Logic
  • Rules
  • Semantic Nets
  • XML
  • Bayesian Nets
• The key point here is that the way knowledge is represented is a crucial aspect of any system and particularly important for an AI system.
• Modified checker board example.
• Tractability vs expressiveness
• Feature selection is important.
• Symbolic vs. subsymbolic vs. statistical