Supervised and Unsupervised Transfer Learning for Question Answering
This addresses the challenge of limited labeled data in QA for researchers and practitioners, though it is incremental as it applies known transfer learning methods to QA.
The paper tackles the problem of applying transfer learning to question answering (QA), showing that a simple technique from MovieQA significantly improves performance on TOEFL and MCTest datasets, with one model achieving state-of-the-art results, including a 7% improvement on TOEFL.
Although transfer learning has been shown to be successful for tasks like object and speech recognition, its applicability to question answering (QA) has yet to be well-studied. In this paper, we conduct extensive experiments to investigate the transferability of knowledge learned from a source QA dataset to a target dataset using two QA models. The performance of both models on a TOEFL listening comprehension test (Tseng et al., 2016) and MCTest (Richardson et al., 2013) is significantly improved via a simple transfer learning technique from MovieQA (Tapaswi et al., 2016). In particular, one of the models achieves the state-of-the-art on all target datasets; for the TOEFL listening comprehension test, it outperforms the previous best model by 7%. Finally, we show that transfer learning is helpful even in unsupervised scenarios when correct answers for target QA dataset examples are not available.