This post is a distractive example of not working, despite the fact that it takes work to do this.
I say distractive because I should be doing something else, and that something else is big and important to me and difficult and terrifying. So this blog is easy.
I hope that by starting something easy (easier?) to write, I will get on a roll or into the practice of spewing my thoughts onto the page. If you are reading this and you go go back and look at my previous, sporadic posts, you may notice a common thread. That thread is not in the content, but in the amount of push they had on my thoughts towards sharing them with the world.
The big work is my PhD dissertation. I claim that computers could model humour appreciation. That is, a computer program could express some "reaction" to something humorous by following a version of the way humans do. I'm not claiming it will have MY sense of humour, but rather it would have ITS sense of humour. Of course this would be based on how it is "trained" and how it stores and retrieves its context.
That's what we do: we learn a bunch about the world ("training and context"), we experience something surprising (unexpected in some way), we resolve that surprise in a safe manner (it's not going to kill us...), and then based on that surprise, the time it takes to resolve, and additional context we will react (laugh, snicker, snort, eye-roll, groan, ignore, etc.).
This is a lot! It is difficult. It is hard (in both the conventional and the computational sense). It has even been said that humour is NP-Complete. I don't know if I could prove this, but that would be a perfect outcome of this project. Doesn't solve humour... but at least it sets the level of challenge and gives us some ways to make accommodations and compromises to approximate the solutions.
All things in, though. I think this is enough effort on this for today. I will now move on to the big work. And come back in a day or two with some distilation of that big work in progress... maybe.
*sigh*
Brains.
No comments:
Post a Comment