French invent sarcasm detector


#1

Well isn't that just lovely. [BBC] READ THE REST


#2

CF


#3

Yes, it's made of cast irony.


#4

Oooh, a sarcasm detector. Now that's a REALLY useful invention.


#5

I can't see how this could possibly fail.


#6

I am sure this works to perfection.


#7

Did they just build a big arrow with the word "sarcasm" on it and point it at the internet?


#8

If it's French, I don't see how it could distinguish the signal from the background noise.


#9

I still say that a Sarcasm tag, along with Bold and Italic is sorely needed in HTML. As it is I just have to put up "Attention all extremely stupid people: The following is an example of Sarcasm. There will be an [End of Sarcasm] tag when the sarcasm is over. If you do not know what sarcasm is follow the above link: new vistas in discourse will open up before you." notices a lot.


#10

Why not put in a smirk instead?


#11

I suspect that the people who don't detect sarcasm in writing generally can't detect it in real life, either. They just think they can, because we can all at least detect ostentatious sarcasm, and real life is generally quieter for the failure to detect sarcasm, whereas online gets louder for it.


#12

It's a sarcasm detector, not a narcissism detector


#13

In the linked article, they seem to be bragging about 80% accuracy. I don't think they understand statistics very well. That's way too inaccurate to be useful.


#15

It's a sarcasm detector, not a narcissism detector

Hey, come on. Vichy was a long time ago.

Oh wait... narcissism! I thought you wrote something else. My bad.


#16

I wonder how the test was done, i.e. if the sample was 50% sarcasm, or the more realistic 1-5%. 80% accuracy on 50% frequency sample can mean a 10% false positive rate (i.e. 10 phrases, five sarcastic phrases were all true positives, four non-sarcastic phrases were true negatives, and one non-sarcastic phrase was a false positive).

If you use a 10% false positive rate test against a 1% frequency sample you get 90% false positive hits within your positive hits (out of a million phrases you'll get eleven thousand positives, out of which ten thousand will be false positives and only one thousand true positives).


#17

This topic was automatically closed after 5 days. New replies are no longer allowed.