Generating Followup Questions for Interpretable Multi-hop Question Answering
This work addresses interpretability in multi-hop question answering for AI researchers, though it is incremental as it builds on existing neural question generation methods.
The paper tackles the problem of answering open domain multi-hop questions by proposing a framework that generates interpretable followup questions, using a pointer-generator network trained on HotpotQA data to achieve competitive performance on two-hop bridge questions.
We propose a framework for answering open domain multi-hop questions in which partial information is read and used to generate followup questions, to finally be answered by a pretrained single-hop answer extractor. This framework makes each hop interpretable, and makes the retrieval associated with later hops as flexible and specific as for the first hop. As a first instantiation of this framework, we train a pointer-generator network to predict followup questions based on the question and partial information. This provides a novel application of a neural question generation network, which is applied to give weak ground truth single-hop followup questions based on the final answers and their supporting facts. Learning to generate followup questions that select the relevant answer spans against downstream supporting facts, while avoiding distracting premises, poses an exciting semantic challenge for text generation. We present an evaluation using the two-hop bridge questions of HotpotQA.