s-nlp/TextGraphs17-shared-task-dataset

We present a dataset for graph-based question answering. The dataset consists of <question; candidate answer> pairs. For each candidate, we present a graph that is obtained by finding the shortest path between named entities mentioned in a question and a candidate answer. As a knowledge graph, we adopted Wikidata. Our dataset has the following fields: sample_id - an identifier for <question, candidate answer>; question - question text; questionEntity - comma-separated list of names (textual… See the full description on the dataset page: https://huggingface.co/datasets/s-nlp/TextGraphs17-shared-task-dataset.