Computer Classes -- Pevensey 2D12

You will take part in one 50 minute computer class per week. starting in week 2, and work in pairs. Working on pairs maximises the chance that you will learn something useful and (for one of the classes) reduces the load on the system thus giving better computer response time. This should be timed to occur (for each of you) earlier in the week than the seminar. There will be material to study prior to the computer class (available on the course website) and participants will be assumed to have studied it. During the class you will run a variety of programs to observe their behaviour. You will not learn to program, but you will vary program parameters (values) to see how they affect the program's behaviour. Note that not all the computer class handouts on the website have yet been updated from last year's versions.

Week 2 -- Introduction (as html, as pdf)

1. Controlling a blocks world by typing questions and commands. Aim: to observe the various stages involved in this model of the transitions from text input to simulated robot action. (Graphic version of TEACH * MSBLOCKS)
2. Running a simulation of children's subtraction behaviour. Aim: to observe in outline how different subtraction errors can be modelled. (TEACH * BUGGY)

Week 3 -- Production Systems (as html, as pdf)

Examining in detail how a production system can be used to model children's subtraction behaviour. Trying out the system with different collections of rules on different subtraction sums.
(revamped graphic version of TEACH * SUBTRACT)

Week 4 -- Neural Networks (as html, as pdf)

Examining the behaviour of a word-recognition neural network (David Young's neural network programme).

Week 5 -- Behaviour-based Modelling (as html, as pdf)

Examining a model of sensors and effectors in a dynamic simulated robot world. (Chris Thornton's Java-based BugWorks programme).

Depending on which group you are allocated to, your class will be either on Tuesdays or Fridays each week, starting in Week 2 -- check on the First Year Noticeboard or the course website. The cohort is divided across 6 computer class groups of roughly 16 students per group.