This is the dataset for the paper:

"Van-Thuy Phi, Joan Santoso, Van-Hien Tran, Hiroyuki Shindo, Masashi Shimbo, and Yuji Matsumoto. Distant Supervision for Relation Extraction via Piecewise Attention and Bag-Level Contextual Inference"

• We provide an annotated dataset of 5,863 sentences, which is checked by annotators for false positive examples. Our dataset can be found in "dataset.txt".

• 1,575 of 5,863 sentences (26.86%) are judged as false positive by three annotators. For 88 sentences (1.5%) for which the two annotators cannot reach an agreement, another participant is involved in the decision-making process. Please check the file "sentences_checked_by_3rd_annotator.txt" for details.

Results in the paper: https://www.dropbox.com/s/c6saejcre9lowhi/models.zip?dl=0

Please CITE the above paper whenever this dataset is used to produce published results or incorporated into other software.