Pilot data annotations for the NLPContributions Model.
The repository is organized as follows:
README.md
[task-name-folder]/ # machine-translation, named-entity-recognition, question-answering, relation-classification, text-classification
├── [article-counter-folder]/ # ranges from 0 to 9 since we annotated 10 articles per task
│ └── research-problem.json # `research problem` mandatory information unit in json format
│ └── model.json # `model` information unit in json format; in some articles it is called `approach`
│ └── ... # there are 8 main information units in all and each article may be annotated by 3 or 6
│ └── triples/ # the folder containing information unit triples one per line
│ │ └── research-problem.txt # `research problem` triples (one research problem statement per line)
│ │ └── model.txt # `model` triples (one statement per line)
│ │ └── ... # there are 8 main information units in all and each article may be annotated by 3 or 6
│ └── ... # there are ten articles annotated for each task, so this repeats nine more times
└── ... # there are five tasks selected overall, so this repeats four more times
Coming soon...