Intelligent Systems - Review Questions
These sets of questions will grow as the course progresses. They arenot intended to be sample exam questions nor are they meant to indicate the scope of knowledge you should have aquired during lectures and tutorials. Many of them have one line answers but this isn't the point either. They are meant to help you think about the stuff you have studied and identify where you need to go back through lecture notes or reach for a book (hopefully both).
Questions about search tools & techniques
- what is meant by search?
- what is the purpose of a legal move generator?
- what's the difference between breadth first & depth first search?
- write a generalised algorithm for a search process
- how can this algorithm be interpreted as (i) breadth first (ii) depth first?
- how can the algorithm be interpreted as best first?
- how can a best 1st search be modified into a heuristic search?
- when is appropriate to use minimax?
- what is the role of a static evaluation function?
- why do you need a static evaluation function?
- how does alpha-beta pruning work? -give an example
- outline 3 other considerations for designing game playing search algorithms
- how can state changing operators be specified (eg: for a simple robot world)?
- give an example of the pick-up operator & state your assumptions about the worlds it can operate in
- why is a standard search algorithm inappropriate for all but the simplest of worlds?
- what do you recommend using instead? explain (in detail) how this works & give an example
- outline the structure of simple expert system rule
- how are rules used?
- what is meant by forward chaining?
- what's backward chaining? give an example & loosely specify an algorithm
- give a couple of examples explaining why an expert system might need to use both forward and backward chaining
- what are certainty factors? how are they used?
- conflict resolution selects rules from the working set - what's this all about?