Semi-Supervised QA with Generative Domain-Adaptive Nets

Abstract

We study the problem of semi-supervised question answering----utilizing unlabeled text to boost the performance of question answering models. We propose a novel training framework, the Generative Domain-Adaptive Nets. In this framework, we train a generative model to generate questions based on the unlabeled text, and combine model-generated questions with human-generated questions for training question answering models. We develop novel domain adaptation algorithms, based on reinforcement learning, to alleviate the discrepancy between the model-generated data distribution and the human-generated data distribution. Experiments show that our proposed framework obtains substantial improvement from unlabeled text.

arXiv commentAccepted as a long paper at ACL2017

References in corpus (17)

• Sequence to Sequence Learning with Neural Networks
• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• Learning Transferable Features with Deep Adaptation Networks
• Bidirectional Attention Flow for Machine Comprehension
• SQuAD: 100,000+ Questions for Machine Comprehension of Text
• Dual Learning for Machine Translation
• Semi-Supervised Learning with Ladder Networks
• On Using Monolingual Corpora in Neural Machine Translation
• Dynamic Coattention Networks For Question Answering
• Machine Comprehension Using Match-LSTM and Answer Pointer
• Modeling Coverage for Neural Machine Translation
• A Thorough Examination of the CNN/Daily Mail Reading Comprehension Task
• Multi-Perspective Context Matching for Machine Comprehension
• Google's Multilingual Neural Machine Translation System: Enabling Zero-Shot Translation
• Review Networks for Caption Generation
• NewsQA: A Machine Comprehension Dataset
• An Empirical Comparison of Simple Domain Adaptation Methods for Neural Machine Translation

Cited by in corpus (16)

• Unsupervised Data Augmentation for Consistency Training
• Self-training with Noisy Student improves ImageNet classification
• Good Semi-supervised Learning that Requires a Bad GAN
• Question Answering and Question Generation as Dual Tasks
• An Introduction to Image Synthesis with Generative Adversarial Nets
• RACE: Large-scale ReAding Comprehension Dataset From Examinations
• A Joint Model for Question Answering and Question Generation
• A Unified Query-based Generative Model for Question Generation and Question Answering
• Finding Answers from the Word of God: Domain Adaptation for Neural Networks in Biblical Question Answering
• Explicit Utilization of General Knowledge in Machine Reading Comprehension
• Two-Stage Synthesis Networks for Transfer Learning in Machine Comprehension
• Data Augmentation for Neural Online Chat Response Selection
• Question Generation from SQL Queries Improves Neural Semantic Parsing
• Evaluating for Diversity in Question Generation over Text
• A Study of the Tasks and Models in Machine Reading Comprehension
• Multi-Perspective Neural Architecture for Recommendation System