The nine types of intelligence
In the beginning of class, Mr Farrell had us find stuff on the nine types of intelligence,
by Howard Gardner. This suggests that the types of intelligence are those
that can be applied to different topics, rather than the difference in how they function.
An example of this is naturalist intelligence, which is regarding sense about nature.
This does not discuss how it works, but what it is.
Alan Turing
We talked some about Alan Turing, a "somewhat" famous computer scientist
who worked on some of the code breaking machines in ww2, and thought of a Turing machine,
which is basically a CPU.
Chat bot
We looked at a chatterbot (I went with Cleverbot, because that is a fairly well known one.)
and we thought some about how realistic conversation
could hold. I have some screenshots of my experiences below:
We are now working on our own chatterbot:
This is my code for a chatterbot. What it does is: First. It asks you your name, then, it asks you your favorite food, and if It hasn't heard it before, it declares it facinating, and adds it to the list of stuff it has heard before. Otherwise, it just dismisses it as boring. It tehn asks you if you have any questions for it, and looks for keywords and patterns, and if none of those are triggered, it goes to wolfram alpha for the answer.
Turing test
We did turing tests on thursday, where one person went outside with the TA, with a G+ chat to the computer hooked up to the smart board, then the TA picked whether the human or their chatterbot would be represented. The class then typed stuff in to the chat, which then went through either the human or chatbot, and the response was put back through to the chat. After a few back and forths, the class voted on whether or not they thought it was human/robot. Alex got a perfect score, because it kept giving syntax errors and people thought it was him trolling.
Watson
Today, friday, we saw a documentary (on youtube of course) about watson winning Jeprody. We saw that watson used something called "Machine Learning" that allowed it to observe patterns in questions and the correct answers. It then drew from a huge database, to find what the correct answer is, based off of what it thinks the parts of the question mean, here is a diagram (because I don't understand ho it works well enough to explain it to you):
We saw war games in between break and lunch, where there was an AI that didn't realize that in a nuclear war, there are no winners. Some kid hacks into the AI and begins its quest for nuclear war, but the AI is stopped at the last moment(of course).