Expert Systems & other important AI terms

Expert Systems & other important AI terms

• ES (Expert System) – A computer program designed to model the problem solving ability of a human expert
• ESDLC
  1. Feasibility study
  2. Rapid prototyping
  3. Alpha system (in-house verification)
  4. Beta system (tested by users)
  5. Maintenance and evolution
• Linear model
  1. Planning
     1. Feasibility assessment
     2. Resource allocation
     3. Task phasing and scheduling
     4. Requirements analysis
  2. Knowledge acquisition and analysis
     • Direct Methods
     • Indirect methods

     main steps in this phase are

     • Knowledge acquisition from expert
     • Define knowledge acquisition strategy (consider various options)
     • Identify concrete knowledge elements
     • Group and classify knowledge. Develop hierarchical representation where possible
     • Identify knowledge source, i.e. expert in the domain

     Techniques

     • Knowledge elicitation by interview
     • Brainstorming session with one or more experts. Try to introduce some structure to this session by defining the problem at hand, prompting for ideas and looking for converging lines of thought
     • Electronic brainstorming
     • On-site observation
     • Documented organizational expertise, e.g. troubleshooting manuals
  3. Knowledge design
  4. Code
  5. Knowledge verification
  6. System evaluation
• computer vision – a subfield of Artificial Intelligence
  help us make machines that can “understand” images and videos
• Inference mechanisms
  • Forward Chaining
  • Backrward Chaining
• MSE (Mean Squared Error)
• Inductive learning is based on the knowledge that if something happens a lot it is likely to be generally true
• 2 main functions of brain in human system
  1. Introspection – that is trying to catch out own thoughts as they go by
  2. Psychological Experiments – that concern with the study of science of mental life
• ANNs (Artificial Neural Networks) – a new learning paradigm which takes its roots from biology
• clustering Algorithm
  • SOM (Self-organizing maps)
  • k-means
  • linear vector quantization
  • Density based data analysis
• Types of rules
  • Relationship
  • Recommendation
  • Directive
  • Variable Rules
  • Uncertain Rules
  • Meta Rules
• Knowledge Base
  • Problem facts, rules
  • Concepts
  • Relationships
• brain – a collection of about 100 billion interconnected neurons
• axon hillock – part of the soma that does concern itself with the signal
• multilayer perceptrons – most basic artificial neural networks
• Fuzzy logic – a superset of conventional (Boolean) logic
• Five parts of the fuzzy inference process
  1. Fuzzification of the input variables
  2. Application of fuzzy operator in the antecedent (premises)
  3. Implication from antecedent to consequent
  4. Aggregation of consequents across the rules
  5. Defuzzification of output
• Softcomputing
  • genetic fuzzy (genetic algorithms – fuzzy systems)
  • neuro-fuzzy systems (Neural Networks – fuzzy systems)
  • neuro-genetic systems (Genetic algorithms – Neural Networks)
• CNF (conjunctive normal form)
• GA (Genetic Algorithms)
• Instance space
• Concept space
• ID3 – interactive dichotomizer