CALBC
From Web4WeB - Portal on Semantic web Technologies
Collaborative annotation of a large biomedical corpus
Project Description
While it is universally acknowledged that annotating scientific literature with formal knowledge would make scientists more productive, these annotations are today very expensive to produce. CALBC's goal is to demonstrate that it is possible to bootstrap automatic (i.e. inexpensive) annotations of satisfactory quality by having a large number of annotation programs all annotate the same corpus in several iterations and improving their accuracy over time by learning from each other's strengths and mistakes. In order to do this CALBC will prepare an appropriately representative corpus of biomedical literature and construct a web based system that would allow any developer of biomedical text-mining applications to submit their annotation of the corpus, determine how this submission diverges from all the others and exploit this information to improve its performance. The final corpus, together with its final annotation (based on a consensus from all the individual submissions) will be made publicly available.
Project Homepage: cordis.europa.eu
| Description | While it is universally acknowledged that … While it is universally acknowledged that annotating scientific literature with formal knowledge would make scientists more productive, these annotations are today very expensive to produce. CALBC's goal is to demonstrate that it is possible to bootstrap automatic (i.e. inexpensive) annotations of satisfactory quality by having a large number of annotation programs all annotate the same corpus in several iterations and improving their accuracy over time by learning from each other's strengths and mistakes. In order to do this CALBC will prepare an appropriately representative corpus of biomedical literature and construct a web based system that would allow any developer of biomedical text-mining applications to submit their annotation of the corpus, determine how this submission diverges from all the others and exploit this information to improve its performance. The final corpus, together with its final annotation (based on a consensus from all the individual submissions) will be made publicly available. ssions) will be made publicly available. |
| HasStartdate | 1 January 2009 + |
| Homepage | http://cordis.europa.eu/ + |
| ProjectStatus | Execution + |
| Subtitle | Collaborative annotation of a large biomedical corpus + |
