| List of files: | |
| ReadMe.txt - This file | |
| doc/WebQSP.pdf - The main document that details the usage, format and other specifics about this WebQuestionsSP dataset | |
| doc/LabelingInstructions.pdf - The annotation guidelines | |
| data/WebQSP.train.json - The training set of the primary dataset | |
| data/WebQSP.test.json - The testing set of the primary dataset | |
| data/WebQSP.train.partial.json - Partial annotations to bad or descriptive questions in the original training set | |
| data/WebQSP.test.partial.json - Partial annotations to bad or descriptive questions in the original testing set | |
| eval/eval.py - The evaluation script in Python | |
| eval/Pred.sem.json - Output of the STAGG system trained using the full semantic parses | |
| For evaluation, please use "WebQSP.test.json". Detailed descriptions of the roles of WebQSP.[train|test].[partial|_].json can be found in "WebQSP.pdf". | |
| Usage of the evaluation script: | |
| python eval.py goldData predAnswers | |
| $ python eval/eval.py data/WebQSP.test.json eval/Pred.sem.json | |
| Number of questions: 1639 | |
| Average precision over questions: 0.709 | |
| Average recall over questions: 0.803 | |
| Average f1 over questions (accuracy): 0.717 | |
| F1 of average recall and average precision: 0.753 | |
| True accuracy (ratio of questions answered exactly correctly): 0.639 | |