Learning to Collaborate for Question Answering and Asking

By Duyu Tang, Nan Duan, Zhao Yan, Zhirui Zhang, Yibo Sun, Shujie Liu, Yuanhua Lv, M. Zhou. 2018

51 citations recorded by the source corpus.

Question answering (QA) and question generation (QG) are closely related tasks that could improve each other; however, the connection of these two tasks is not well explored in literature. In this paper, we give a systematic study that seeks to leverage the connection to improve both QA and QG. We present a training algorithm that generalizes both Generative Adversarial Network (GAN) and Generative Domain-Adaptive Nets (GDAN) under the question answering scenario. The two key ideas are improving the QG model with QA through incorporating additional QA-specific signal as the loss function, and improving the QA model with QG through adding artificially generated training instances. We conduct experiments on both document based and knowledge based question answering tasks. We have two main findings. Firstly, the performance of a QG model (e.g in terms of BLEU score) could be easily improved by a QA model via policy gradient. Secondly, directly applying GAN that regards all the generated questions as negative instances could not improve the accuracy of the QA model. Learning when to regard generated questions as positive instances could bring performance boost.

Subfields: Artificial Intelligence, Natural Language Processing, Reinforcement Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Reinforcement Learning

Citations in the corpus

Addressing Semantic Drift in Question Generation for Semi-Supervised Question Answering (2019, methodology): “Yang et al. (2017) and Tang et al. (2018) followed the GAN (Good-fellow et al., 2014) paradigm, taking QG as a generator and QA as a discriminator, to utilize unlabeled data.”
Review-based Question Generation with Adaptive Instance Transfer and Augmentation (2019, background): “…with its dual task, i.e., reading comprehension or question answering with various motivations, such as improving auxiliary task performance (Duan et al., 2017; Yang et al., 2017; Golub et al., 2017), collaborating QA and QG model (Tang et al., 2018, 2017), and uniﬁed learning (Xiao et al., 2018).”
Semantic Parsing with Dual Learning (2019, methodology): “The idea of dual learning has been applied in various tasks (Xia et al., 2017), such as Question Answering/Generation (Tang et al., 2017, 2018), Image-to-Image Translation (Yi et al., 2017) and Open-domain Information Extraction/Narration (Sun et al., 2018).”
COOT: Cooperative Hierarchical Transformer for Video-Text Representation Learning (2020, methodology): “The assumption of cyclical structure has been used in various works [83, 84, 85].”
Learning to Generate Questions by LearningWhat not to Generate (2019, background): “y a difficulty estimator, and generate questions on different difficulty levels. [33] enriches the model with both answer embedding and the relative distance between the context words and the answer. [34, 35] model Table 2: Evaluation results of different models on SQuAD dataset. Model BLEU-1 BLEU-2”
Improving Neural Question Generation using Answer Separation (2018, methodology): “It is also used as a way to improve question answering (QA) systems (Duan et al. 2017; Tang et al. 2017; Tang et al. 2018) or to engage chatbots to start and continue a conversation (Mostafazadeh et al. 2016).”