CPSC 433

We organized into groups and talked about the course in general. If you were not able to attend the first lab please contact me for group assignment.

We discussed ELIZA, a replacement psychiatrist. More importantly we discussed how we can process natural language to seem like we are paying attention.

I recomend looking at a couple chatterbots:

• A.L.I.C.E - a `bot' with more engineering effort.
• JabberWacky - a `bot' which learns from the people who interact with it.
• CleverBot - a cousin to JabberWacky that also learns from people.

There are also some implementations of ELIZA herself which are worth looking at.

• Eliza Test is a Java implementation of ELIZA from Charles Hayden. It's faithful to the original implementation baring some changes to how the script files are written.
• The ELIZA built into Emacs (not just x-emacs as I mentioned) is entertaining. I suspect it has some advancements beyond the original implementation. You can invoke using M-x doctor.

If you're interested, check out the original paper by Joseph Weizenbaum. The link should work on campus, from off campus you will need to search through the U of C Library.

In this tutorial and the 2 following we covered the high level definitions of search systems. We illustrated them using a puzzle as an example. These slides, Model 1 and Model 2 provide examples (with slightly better explanations than in tutorial) of the operation of the different models and controls we described.

We continued the puzzle example.

We continued the puzzle example.

We started to cover Evolutionary Algorithms (or Genetic Algorithms) as an example of set based search. We looked at how we describe individuls and genetic operators as facts and extension rules and at how we define the functions of our search control.

We will do an example of an EA system solving the 0/1 knapsack problem. These slides include a further overview of EAs and another example run of this system.

If there is time we will also briefly introduce Branch and Bound Trees.

We looked at formal definitions for Branch and Bound trees as an example of and-trees. We did one example for a search control, that and a second search control and example are on these slides .

My slides from previous years on first order logic and the conversion from FOL to clauses for resolution.

In an effort to make the XML stuff slightly clearer I put together some new slides. Also take a look at XML Files and W3 Schools. Given that these are very new slides I probably still have some errors in them, please let me know if you spot any. I am also happy to answer any questions you may have.

A few links on Conceptual Dependency. One simple overview is David Marshal's (although his graphics do not seem to render properly), another is Anthony Francis's. You might also be interested in the original paper on a conceptual dependency parser which explains several of the motivations behind CD. Finally there are some other examples available on this semantic network overview.

For an example of perceptrons please consult my perceptron slides