[Corpora-List] EmoText - Software for opinion mining and lexical affect sensing

[Yannick Versley]

>
> The system relies on scientific findings in my phd thesis; it works and
> the results are not that bad. They could be better and I also describe in
> the thesis how they can be improved. It's up to you, you can ignore them
> and loose time.

It's a common fallacy that we in computational linguistics (maybe in all
applicable science in general)
tend towards, that if we made scientific progress, what we made must
obviously be good for something.

This is even more valid for tasks that do have a strong
application-oriented motivation, such as
sentiment analysis, and as a researcher you have a responsability towards
society not to live
in too bad a delusion about the usability of "exciting" work.
(Most researchers cultivate a bit of reality distortion to be able to get
their work publish, but let's
not think that having this kind of delusion is good for society).

Turning scientific research into a real improvement (for your customers, or
for society in general)
is a long way from the actual research, and I do think it's laudable to
build demonstrators that
allow people to inspect the current state of the art, but the remaining
(larger) part of the long way
means that you have to step outside the box of academic evaluation and
assess the whole
context of usage (see, e.g., the "good uses for bad MT" paper that I can't
seem to find).

-Yannick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20111217/8ae03d88/attachment.htm>
-------------- next part --------------
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora