This code implements a word_overlap retrieval baseline. given a question and a file consisting all the knowledge, e.g. list of all table row sentences.
In order to score the retrieval model, picks a random list of question, answer pairs (e.g. from AristoQuiz dataset) and it checks if the answer exists in the topk retrieved sentences.
btw, the code has been checked by Joel
This code implements a word_overlap retrieval baseline. given a question and a file consisting all the knowledge, e.g. list of all table row sentences. In order to score the retrieval model, picks a random list of question, answer pairs (e.g. from AristoQuiz dataset) and it checks if the answer exists in the topk retrieved sentences. btw, the code has been checked by Joel