Lab Exercise 10: Humor and Artificial Intelligence
While most of the topics we discussed in this class are intended to be taken seriously, there is also room for humor and fun in AI. In this exercise, I want you to find one example of humor in Artificial Intelligence. This can be intentional or unintentional (e.g. some conversation attempts by chatbots), cartoons, jokes, short stories, events, movie scenes, game episodes, or something else. And there are even people who are writing serious research articles about this, trying to figure out if computers can understand word plays or jokes.
Instructions
Submit your AI humor piece to the "Lab 10" discussion board on Blackboard. If the item is available on the Web, a link is fine, although there is the risk that the link may not last. In order to avoid duplicate submissions, we'll follow a first-come, first-post rule: If somebody else posted the same piece before you did, you'll have to find another one.
The topic should be related to Artificial Intelligence; generic geek humor, nerd jokes, or computer cartoons without an AI link may not qualify. You can point out the connection to AI if it is not immediately recognizable. But as with humor in general, if there's a need to explain it, most likely it's not that funny. The one exception that I have no problems accepting here is jokes in Lojban ...
Please avoid submissions that may be offensive to others. If you're in doubt, check with your team mates or me before you post it.
If it is not possible to submit your piece electronically, you can give me a paper copy. I'd prefer electronic submission, however, so that others can enjoy it as well.
Grading Procedure
Your submissions will be used as percepts for a simple reflux agent. That agent serves as performance element in a complex learning system that incorporates neural networks, genetic algorithms, and fuzzy logic. It attempts to improve the condition-action rules so that the agent can learn to understand humor it previously had not been exposed to. Immediate feedback may be provided to the agent through comments and reactions of other agents in the vicinity in response to the agent's behavior when reading the submissions. After processing all submissions, the trained agent will assign scores to the individual entries. If the agent fails to "get" it, it will assign a random score, biased by the number of words recognizable as related to AI.