Groups
Groups
Sign in
Groups
Groups
BEA 2019 Shared Task: Grammatical Error Correction
Conversations
About
Send feedback
Help
BEA 2019 Shared Task: Grammatical Error Correction
Contact owners and managers
1–30 of 34
### Welcome to the discussion forum for the BEA 2019 Shared Task on Grammatical Error Correction! ###
This forum is the best place to receive updates about the shared task and post questions.
You should definitely subscribe to this group if you are intending to participate in the shared task!
BEA 2019 Shared Task Organisers
Website:
https://www.cl.cam.
ac.uk/research/nl/bea2019st/
Mark all as read
Report group
0 selected
BEA 2019 Shared Task Organisers
7/21/19
BEA2019 Shared Task - Codalab Outage
Dear Shared Task participants, You may or may not have noticed that Codalab suffered a major
unread,
BEA2019 Shared Task - Codalab Outage
Dear Shared Task participants, You may or may not have noticed that Codalab suffered a major
7/21/19
BEA 2019 Shared Task Organisers
4/4/19
BEA2019 Shared Task - Results released!
Dear all, We have just released the shared task results! You can find them on Codalab under the
unread,
BEA2019 Shared Task - Results released!
Dear all, We have just released the shared task results! You can find them on Codalab under the
4/4/19
BEA 2019 Shared Task Organisers
2/14/19
***Important*** Shared Task FAQs
This thread contains a summary of some of the most important questions asked about the shared task so
unread,locked,
***Important*** Shared Task FAQs
This thread contains a summary of some of the most important questions asked about the shared task so
2/14/19
Ava Morgan
12/8/25
RICHARD IS COMING AGAIN
Hey neighbors, Richard is coming to my house to clean my Ducts and vent. I had his service last year
unread,
RICHARD IS COMING AGAIN
Hey neighbors, Richard is coming to my house to clean my Ducts and vent. I had his service last year
12/8/25
Hiruy Getachew
5/6/24
requesting the code
can you give me the full python code for this grammatical error correction?
unread,
requesting the code
can you give me the full python code for this grammatical error correction?
5/6/24
BEA 2019 Shared Task Organisers
7/11/22
BEA 2019 Codalab Platform Update
Dear all, As organisers of the BEA 2019 Shared Task on Grammatical Error Correction, we wish to
unread,
BEA 2019 Codalab Platform Update
Dear all, As organisers of the BEA 2019 Shared Task on Grammatical Error Correction, we wish to
7/11/22
BEA 2019 Shared Task Organisers
5/24/19
BEA19 - Shared Task Reference
Dear BEA19 shared task participants, This is just a quick message about how to reference the shared
unread,
BEA19 - Shared Task Reference
Dear BEA19 shared task participants, This is just a quick message about how to reference the shared
5/24/19
BEA 2019 Shared Task Organisers
5/18/19
BEA2019 Shared Task - Paper Reviews
Dear BEA2019 Shared Task participants, For those of you who submitted papers, apologies for the delay
unread,
BEA2019 Shared Task - Paper Reviews
Dear BEA2019 Shared Task participants, For those of you who submitted papers, apologies for the delay
5/18/19
BEA 2019 Shared Task Organisers
5/8/19
BEA19 Shared Task Submission Descriptions
Dear Shared Task Participants, First, I'd like to thank all teams that submitted system papers to
unread,
BEA19 Shared Task Submission Descriptions
Dear Shared Task Participants, First, I'd like to thank all teams that submitted system papers to
5/8/19
BEA 2019 Shared Task Organisers
,
leshem....@mail.huji.ac.il
3
4/18/19
BEA19 - System output released + dev request
We'll have to download and share them manually unfortunately. There is a download button on
unread,
BEA19 - System output released + dev request
We'll have to download and share them manually unfortunately. There is a download button on
4/18/19
Abhijeet Awasthi
, …
leshem....@mail.huji.ac.il
3
4/15/19
Release of system outputs (predicted correct sentences) of participating systems
Hey Chris, Is it possible to gather also the results on the dev (and perhaps train) sets as well for
unread,
Release of system outputs (predicted correct sentences) of participating systems
Hey Chris, Is it possible to gather also the results on the dev (and perhaps train) sets as well for
4/15/19
Felix Stahlberg
,
BEA 2019 Shared Task
2
4/12/19
System description papers
Hi Felix, Another participant actually asked the same question in a private email so it's
unread,
System description papers
Hi Felix, Another participant actually asked the same question in a private email so it's
4/12/19
Xiaoyun Shao
,
BEA 2019 Shared Task Organisers
2
4/3/19
Which software can be used to analyse this corpus?
Dear Vikki, All the corpora provided in the shared task are plain text files (m2 format) so there is
unread,
Which software can be used to analyse this corpus?
Dear Vikki, All the corpora provided in the shared task are plain text files (m2 format) so there is
4/3/19
BEA 2019 Shared Task Organisers
,
tmu
3
4/1/19
Test data released! - 25/3/19
Hi, There are currently no plans to release the target side of the test data. We are deliberately
unread,
Test data released! - 25/3/19
Hi, There are currently no plans to release the target side of the test data. We are deliberately
4/1/19
Yoav Katz
,
BEA 2019 Shared Task
5
3/29/19
Use of external tools in Restricted Track
Hi Yoav, Even though it's open source, we also have a question about novelty. Since the main goal
unread,
Use of external tools in Restricted Track
Hi Yoav, Even though it's open source, we also have a question about novelty. Since the main goal
3/29/19
awasthia...@gmail.com
,
BEA 2019 Shared Task Organisers
2
3/28/19
Max-Match scorer Vs Span-based correction in ERRANT
Yes. MaxMatch attempts to extract the edits from the hypothesis such that they maximally overlap with
unread,
Max-Match scorer Vs Span-based correction in ERRANT
Yes. MaxMatch attempts to extract the edits from the hypothesis such that they maximally overlap with
3/28/19
wangw...@gmail.com
, …
BEA 2019 Shared Task
6
3/28/19
Max submissions total
Yes, but this is the total limit, *not* per day. If you make your 2 submissions today for a given
unread,
Max submissions total
Yes, but this is the total limit, *not* per day. If you make your 2 submissions today for a given
3/28/19
14484...@qq.com
,
BEA 2019 Shared Task Organisers
2
3/27/19
Same sentences, but not normalize
Welcome to the wonderful world of valid alternative answers! All of those are correct, but that's
unread,
Same sentences, but not normalize
Welcome to the wonderful world of valid alternative answers! All of those are correct, but that's
3/27/19
robin
, …
BEA 2019 Shared Task Organisers
6
3/27/19
About a kind of correction
I'm afraid you're more or less asking us for the correct answer to a particular case so we
unread,
About a kind of correction
I'm afraid you're more or less asking us for the correct answer to a particular case so we
3/27/19
14484...@qq.com
,
BEA 2019 Shared Task Organisers
2
3/23/19
In dev dataset, some lines are not a sentences?
So when we preprocess the data, we split it on newlines on the assumption that a newline ends a
unread,
In dev dataset, some lines are not a sentences?
So when we preprocess the data, we split it on newlines on the assumption that a newline ends a
3/23/19
Abhijeet Awasthi
,
BEA 2019 Shared Task Organisers
4
3/21/19
Tokenization issues
I guess the bottom line is that yes, there are some tokenisation issues in the data, but everyone
unread,
Tokenization issues
I guess the bottom line is that yes, there are some tokenisation issues in the data, but everyone
3/21/19
robin
,
BEA 2019 Shared Task Organisers
2
3/21/19
About the submission rule
Hi, 1. We will make the test data available on Codalab late on Sunday night, shortly before the test
unread,
About the submission rule
Hi, 1. We will make the test data available on Codalab late on Sunday night, shortly before the test
3/21/19
14484...@qq.com
, …
BEA 2019 Shared Task Organisers
4
3/15/19
the corrected sentences of ABCN.dev.gold.bea19.m2 only get a f0.5 score 86.45 on Codalab platform.
No, the corrected sentences are always the same, it's just the edit spans that are different; eg
unread,
the corrected sentences of ABCN.dev.gold.bea19.m2 only get a f0.5 score 86.45 on Codalab platform.
No, the corrected sentences are always the same, it's just the edit spans that are different; eg
3/15/19
writeto...@gmail.com
,
BEA 2019 Shared Task Organisers
2
3/14/19
Alternative correct versions of an incorrect sentence
Yes, good question. We had not announced it yet because annotating is still underway, but we expect
unread,
Alternative correct versions of an incorrect sentence
Yes, good question. We had not announced it yet because annotating is still underway, but we expect
3/14/19
wangw...@gmail.com
,
BEA 2019 Shared Task Organisers
2
3/14/19
Use of Development Data Set
During the development phase, it's probably better not to include it as training data so you can
unread,
Use of Development Data Set
During the development phase, it's probably better not to include it as training data so you can
3/14/19
julia
, …
BEA 2019 Shared Task Organisers
6
3/9/19
Is it allowed to normalize use of apostrophe?
We considered that too, but as we're not the original authors of NUCLE or Lang-8, we didn't
unread,
Is it allowed to normalize use of apostrophe?
We considered that too, but as we're not the original authors of NUCLE or Lang-8, we didn't
3/9/19
robin
,
BEA 2019 Shared Task Organisers
4
3/8/19
Question about the difference between the devset we download and the Codelab's
You are using an old version of the data. Please download v2 from the website. Let us know if you
unread,
Question about the difference between the devset we download and the Codelab's
You are using an old version of the data. Please download v2 from the website. Let us know if you
3/8/19
irina.maslowski
,
BEA 2019 Shared Task Organisers
2
3/7/19
Evaluation of multi-solution results
Hi Irina, The table on the website is mainly an example of ERRANT behaviour. Since we only ask
unread,
Evaluation of multi-solution results
Hi Irina, The table on the website is mainly an example of ERRANT behaviour. Since we only ask
3/7/19
Yoav Katz
, …
BEA 2019 Shared Task Organisers
7
3/6/19
Multisentence context in WI+locness dataset
No. We will only shuffle the test data in terms of essays; eg B essay, C essay, A essay, N essay, A
unread,
Multisentence context in WI+locness dataset
No. We will only shuffle the test data in terms of essays; eg B essay, C essay, A essay, N essay, A
3/6/19
BEA 2019 Shared Task Organisers
3/6/19
BEA19 Shared Task Update - 6/3/2019
Dear BEA19 Shared Task Participants, We have a series of important announcements to make: 1. We have
unread,
BEA19 Shared Task Update - 6/3/2019
Dear BEA19 Shared Task Participants, We have a series of important announcements to make: 1. We have
3/6/19
Xiaoyun Shao
,
BEA 2019 Shared Task Organisers
2
2/27/19
questions about NUCLE corpus
Hi Xiaoyun, The download for NUCLE is in tar bzip2 format. You can open it from the command line in
unread,
questions about NUCLE corpus
Hi Xiaoyun, The download for NUCLE is in tar bzip2 format. You can open it from the command line in
2/27/19
Yoav Katz
,
BEA 2019 Shared Task Organisers
3
2/25/19
Differences when generating M2 for same text
Reading your post again more closely, I realised you were actually asking about gold vs auto
unread,
Differences when generating M2 for same text
Reading your post again more closely, I realised you were actually asking about gold vs auto
2/25/19
14484...@qq.com
,
BEA 2019 Shared Task Organisers
2
2/18/19
About the NUCLE dataset
That is correct. Here are the statistics for all the datasets allowed in the Restricted Track: FCE:
unread,
About the NUCLE dataset
That is correct. Here are the statistics for all the datasets allowed in the Restricted Track: FCE:
2/18/19