Question Answering through Transfer Learning from Large Fine-grained Supervision Data
This addresses the problem of enhancing QA performance for researchers and practitioners, but it is incremental as it applies a basic transfer learning technique to existing datasets.
The paper tackled improving question answering by using transfer learning from a large, fine-grained dataset (SQuAD), achieving state-of-the-art results with over 8% improvement on WikiQA and also on SemEval-2016 and an entailment task.
We show that the task of question answering (QA) can significantly benefit from the transfer learning of models trained on a different large, fine-grained QA dataset. We achieve the state of the art in two well-studied QA datasets, WikiQA and SemEval-2016 (Task 3A), through a basic transfer learning technique from SQuAD. For WikiQA, our model outperforms the previous best model by more than 8%. We demonstrate that finer supervision provides better guidance for learning lexical and syntactic information than coarser supervision, through quantitative results and visual analysis. We also show that a similar transfer learning procedure achieves the state of the art on an entailment task.