Skip to Content.
Sympa Menu

illinois-ml-nlp-users - Re: [Illinois-ml-nlp-users] Loading these sentences takes too much time

illinois-ml-nlp-users AT lists.cs.illinois.edu

Subject: Support for users of CCG software closed 7-27-20

List archive

Re: [Illinois-ml-nlp-users] Loading these sentences takes too much time


Chronological Thread 
  • From: Astronaut Guo <astronautguo AT gmail.com>
  • To: Nicholas Rizzolo <rizzolo AT gmail.com>
  • Cc: illinois-ml-nlp-users <illinois-ml-nlp-users AT cs.uiuc.edu>
  • Subject: Re: [Illinois-ml-nlp-users] Loading these sentences takes too much time
  • Date: Sat, 23 Oct 2010 15:07:41 +0800
  • List-archive: <http://lists.cs.uiuc.edu/pipermail/illinois-ml-nlp-users>
  • List-id: Support for users of CCG software <illinois-ml-nlp-users.cs.uiuc.edu>

Hi Nick,

Thank you very much for your reply.
Now my solution is to add a capital letter in front of the text where the first letter is not a capital letter.
I wonder why such sentences is OK in the online demo. Is there any other solutions?

Thanks a lot.

--Yuhang


On Sat, Oct 23, 2010 at 12:31 AM, Nicholas Rizzolo <rizzolo AT gmail.com> wrote:
Hi Yuhang,

Thanks very much for the bug report.  It turns out the problem was
with a regular _expression_ in LBJ's sentence splitter, which seemingly
takes exponentially more time as the first capital letter of the text
gets further from the start of the text.  (I had no idea such behavior
was possible!)  Anyway, this bug will be fixed when the next version
of LBJ is released, which should be soon.  Stay tuned to this mailing
list for an announcement when it is released.  At that time, you'll
need to download both LBJ and the coref package again.

Thanks,
 - Nick


On Thu, Oct 21, 2010 at 9:39 AM, Astronaut Guo <astronautguo AT gmail.com> wrote:
> Hi,
> When I run CorefPlainText from the Coreference Package I find it takes too
> much time (at least 2 hours no result and I stopped it) to handle the text
> below.
> --------------------------
> 9 suspicious packages planted around Boston in backfired marketing ploy by
> TV network BOSTON 2007-01-31 22:07:50 UTC.
> At least nine electronic devices, planted at bridges and other parts of
> Boston as part of a marketing campaign for a late-night cartoon, threw a
> scare into the city Wednesday.
> Highways, bridges and a section of the Charles River were shut down and bomb
> squads were sent in before authorities declared the devices were harmless.
> "It's a hoax -- and it's not funny," said Massachusetts Gov. Deval Patrick.
> Turner Broadcasting, parent company of Cartoon Network, said the devices,
> which consisted of magnetic, blinking lights, were part of a promotion for
> the TV show "Aqua Teen Hunger Force."
> --------------------------
> I tracked the codes and found that the program can't get through in line 88
> (Doc doc = loader.loadDoc(fullText);), CorefPlainText.class.
> However this text is OK in the online demo. I'm not sure if this is a bug.
> Does anyone know where is the problem?
> Thank you.
> --Yuhang
>
> _______________________________________________
> illinois-ml-nlp-users mailing list
> illinois-ml-nlp-users AT cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/illinois-ml-nlp-users
>
>




Archive powered by MHonArc 2.6.16.

Top of Page