|Originally posted by trancaholic |
@Andrew: Don't believe it for a second. As I may have mentioned before, my job is research within computer science, with the focus area being - yes - probabilistic AI and decision theory. I took a look at the professors website, and was instantly alarmed by him being a professor of sociology. Now researchers in the humanities rarely come up with ideas within computer science that takes the entire field years into the future. Just as computer scientists don't come up with theories on human behaviour that blows all current theories in the social sciences out of the water. And believe you me, we are very far from having methods that can grade a text automatically. Hell, we cannot even recognize spam automatically too a satisfying degree.
Now the professor is *also* an adjunct in computer science, so not wishing to make a too rash judgement I tried to get a peek at the actual reasoning behind the computer programs. This was not available (that is sooo not scientific to keep that secret), but the professor is from a department of "applied expert systems". In science, the "applied" bit usually means that the people there don't have a clue what they're working with, but considers it a black box. This suspicion of mine ties in nicely with the fact that "expert systems" themselves have been abandoned by the AI community for years, as they are incapable of generalizing beyond the examples they have been constructed to recognize, and are computationally speaking highly inefficient. My guess is that the described computer program searches for some sequences of words in the text and increases a grade counter whenever it finds one such sequence. If the student chooses to answer in an unusual way, the program will output a low grade. Simple as that.
Oh, btw. all the programs the professor has constructed has a ™ added to their name ("Data Collection Selection™", "Measurement & Scaling Strategist™", etc.). How cheesy and unscientific is that!
What type of research do you do exactly in decision theory?