search

laito / cleartk

Automatically exported from code.google.com/p/cleartk
0 stars 0 forks source link

write paper for LREC 2014 #363

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
We should write a paper for the 3rd UIMA@GSCL Workshop:

https://docs.google.com/document/d/18BfopkrOuGlhqybPf4z25_FYbA5hal7YvS7hSod4_-U/
pub

The submission deadline is July, 12. Here are the things we discussed today:

== Just give me an annotator ==
CleartkAnnotator is just a JCasAnnotator
Chunking is just a utility object for use in a JCasAnnotator
Features like TF-IDF are in the Annotator, not in the encoder

== Just give me a pipeline ==
New evaluation's train and test methods
Trainable extractors for TF-IDF etc. (not encoders)

== Keep CollectionReaders minimal ==
URICollectionReader

== Don't organize projects by type ==
cleartk-typesystem
cleartk-corpus
cleartk-feature-extractor

== Make annotators type-system agnostic ==
ClearNLP work

== How to get a student to use ClearTK ==
Write the reader and eval, let the student feature-engineer

Original issue reported on code.google.com by steven.b...@gmail.com on 16 Apr 2013 at 7:19

GoogleCodeExporter commented 9 years ago 
I put the above outline into the UIMA@GSCL format and committed that in 
revision 7cbb5ed360e8de465107ffaeba090a2631f40517.

Original comment by steven.b...@gmail.com on 8 May 2013 at 1:03

GoogleCodeExporter commented 9 years ago 
Looks like we're not going to make UIMA@GSCL. We agreed to instead target LREC 
2014 in Reyjkavik:

http://www.lrec-conf.org/lrec2014/

The submission deadline is 15 October 2013 and abstracts are about 1500-2000 
words.

Original comment by steven.b...@gmail.com on 25 Jun 2013 at 11:30

GoogleCodeExporter commented 9 years ago 
Let's think about what kind of survey might be useful to send out to 
cleartk-users that might be worth reporting in this write-up.

Original comment by phi...@ogren.info on 23 Jul 2013 at 3:57

GoogleCodeExporter commented 9 years ago 
We can get download statistics from Maven Central:

http://blog.sonatype.com/people/2010/12/now-available-central-download-statistic
s-for-oss-projects/

After logging in, I can see, for example, that over the last year (14-Aug-2012 
was our first upload to Maven Central) there were 603 unique IP addresses that 
downloaded cleartk-util, 574 that downloaded cleartk-ml, etc.

It's not nearly as nice as Google Analytics - we can't see where in the world 
the IP addresses are from, for example. But at least it's a start.

Original comment by steven.b...@gmail.com on 29 Jul 2013 at 9:44

GoogleCodeExporter commented 9 years ago 
The paper was submitted to LREC on 14 Oct 2013.

Original comment by steven.b...@gmail.com on 25 Oct 2013 at 8:54