BARR 2017
The proposed Biomedical Abbreviation Recognition and Resolution (BARR) track - 2017
has the aim to promote the development and evaluation of biomedical abbreviation identification systems by providing Gold Standard training, development and test corpora manually annotated by domain experts with abbreviation-definition pairs within abstracts of biomedical documents written in Spanish.
BARR track asks participating teams to test processing text systems that are able to detect explicit occurrences of abbreviation-definition pairs.
See more information at the
call for papers
Important tentative dates
-
March 28th, 2017: Release of sample data May 19th, 2017: Release of training data subset1-
May 30th, 2017: Release of Training Subset 2 - June 19th, 2017: Release of Testing Set
- June 26th, 00:00:00 UTC+02:00 Europe/Berlin, 2017: Submission of participant runs
- July 1st, 2017: Working notes submission due (short system description 3-5 pages)
- July 3rd, 2017: Reviews of Working notes sent out to authors.
- July 7th, 2017: Deadline to submit Camera ready revised Working notes
- September 19th, 2017: Workshop at SEPLN 2017
Scientific Advisory Board
- Dr. Saber Ahmad Akhondi (Principle NLP Scientist, Elsevier Content & Innovation)
- Dr. Marius Doornenbal (Chief NLP Scientist, Elsevier Content & Innovation)
- Dr. Georgios Tsatsaronis (Senior NLP Scientist, Elsevier Content & Innovation)
BARR track organizers
- Martin Krallinger, Biological Text Mining Unit (Bio-TeMUC), CNIO, Spain
- Santiago de la Peña, Biological Text Mining Unit (Bio-TeMUC), CNIO, Spain
- Ander Intxaurrondo, Biological Text Mining Unit (Bio-TeMUC), CNIO, Spain
- Jesús Santamaría, Biological Text Mining Unit (Bio-TeMUC), CNIO, Spain
- Jose A. Lopez-Martin, Medical Oncology, Hospital 12 de Octubre, Spain
- Alfonso Valencia, Life Sciences & Computational Biology, BSC, Spain
- Marta Villegas, Life Sciences & Computational Genomics, BSC, Spain
- Anália Lourenço, Next Generation Computer Systems Group, University of Vigo, Spain
- Gael Pérez,Next Generation Computer Systems Group, University of Vigo, Spain
- Martín Pérez, Next Generation Computer Systems Group, University of Vigo, Spain
| Name | Date | Size |
|---|---|---|
| BARR-E background corpus.zip | 2019-05-30 11:16:53 | 65.27MB |
| relations_example.tsv | 2019-05-30 11:16:53 | 375.00B |
| BARR_ibereval_training1_set.zip | 2019-05-30 11:16:52 | 168.59kB |
| BARR_ibereval_training2_set.zip | 2019-05-30 11:16:52 | 311.08kB |
| entities_example.tsv | 2019-05-30 11:16:52 | 468.00B |
| BARR_ibereval_sample_set.zip | 2019-05-30 11:16:52 | 61.34kB |
| BARR_ibereval_training_full_set.zip | 2019-05-30 11:16:52 | 612.82kB |
| BARR_ibereval_relation_evaluation tutorial.pdf | 2019-05-30 11:16:52 | 944.85kB |
| BARR_ibereval_background_set README.txt | 2019-05-30 11:16:52 | 631.00B |
| BARR_ibereval_background1_set.zip | 2019-05-30 11:16:52 | 49.93MB |
| BARR_ibereval_entity_evaluation tutorial.pdf | 2019-05-30 11:16:52 | 1.25MB |
| BARR relation_evaluation tutorial.pdf | 2019-05-30 11:16:51 | 927.38kB |
| BARR_CFP.pdf | 2019-05-30 11:16:51 | 91.04kB |
| BARR_CFP_espanol.pdf | 2019-05-30 11:16:51 | 91.39kB |
| BARR entity_evaluation tutorial.pdf | 2019-05-30 11:16:51 | 1.23MB |
-
Abbreviation/acronym tagger
-
Datasets and lexical resources:
-
Resources of spanish biomedical acronyms and abbreviations:
-
NLP software:
-
Machine Learning:
-
What are the supported communication formats?
-
What programming languages are accepted?
-
Which are the mandatory REST API methods?
-
Which are the recommended browsers?
-
How were the document identifiers generated?
-
Where did the abstracts come from?
-
I discovered records not written in Spanish, will they we included for evaluation?
-
I found that some records have at the end "(AU)", what is this and why was it not annotated?
-
I found that there were some problems with word boundaries, why is that the case?
-
Do I need to recognize all the annotations provided for BARR?
-
I've found that some short forms keep their English aspect in Spanish records and can be linked within them to both their respective English or equivalent Spanish long forms. Should I detect both associations or only the Spanish one?
-
I've found a long form in the abstract, but the short form is not mentioned until later in the abstract, in a different sentence. Should I tag that long form as well?
-
What do we consider as abbreviations? Should we tag proportion units, variables or names of species?
-
Why do we limit the task to long form-short forms (LF-SF)?
-
What data and information will the participants receive?
-
What will the available input formats/files be?
-
How will the training/test phase proceed?
-
Can I adapt, retrain or integrate existing software
-
What methods are allowed to participate?
-
How were the used abstracts selected?
-
What was the background of the curators that prepared the dataset?
-
How were the training, development and test set selected?
-
I have found some missing annotation in the dataset, what should I do?
-
Are the evaluation results anonymous?