Leveraging Context Information for Natural Question Generation

By Linfeng Song, Zhiguo Wang, Wael Hamza, Daniel Gildea. 2018

161 citations recorded by the source corpus.

The task of natural question generation is to generate a corresponding question given the input passage (fact) and answer. It is useful for enlarging the training set of QA systems. Previous work has adopted sequence-to-sequence models that take a passage with an additional bit to indicate answer position as input. However, they do not explicitly model the information between answer and other context within the passage. We propose a model that matches the answer with the passage before generating the question. Experiments show that our model outperforms the existing state of the art using rich features.

Subfields: Artificial Intelligence, Natural Language Processing

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

Citations in the corpus

Addressing Semantic Drift in Question Generation for Semi-Supervised Question Answering (2019, background): “…later works leveraged information from the whole paragraph for QG, either by extracting additional information from the paragraph (Du and Cardie, 2018; Song et al., 2018; Liu et al., 2019) or by directly taking the whole paragraph as input (Zhao et al., 2018; Kim et al., 2018; Sun et al., 2018).”
Learning to Generate Questions by LearningWhat not to Generate (2019, methodology): “nd for the NewsQA dataset, we can only report the baselines with open source code available. •PCFG-Trans [12] is a rule-based system that generates a question based on a given answer word span. •MPQG [31] proposed a Seq2Seq model that matches the answer with the passage before generating the questio”
Evaluating Rewards for Question Generation Models (2019, methodology): “Song et al. (2018) use a modified context encoder based on multiperspective context matching (Wang et al., 2016).”
Low-Resource Generation of Multi-hop Reasoning Questions (2020, background): “Subsequently, we exploit multi-perspective fusion (Song et al., 2018) to grasp the answer-related context in the sentences and strengthen their cross interactions.”
Putting the Horse before the Cart: A Generator-Evaluator Framework for Question Generation from Text (2018, methodology): “(Song et al., 2018) (in an approach referred to as NQG LC hereafter) encoded ground-truth answers and employed bi-directional LSTMs in a Seq2Seq setting.”
Substructure Substitution: Structured Data Augmentation for NLP (2021, methodology): “Relatedly, automatic question generation has been used in data augmentation for question answering (Yang et al., 2017; Song et al., 2018).”
Two Birds, One Stone: A Simple, Unified Model for Text Generation from Structured and Unstructured Data (2019, methodology): “Moreover, we use the vocabulary set released by Song et al. (2018) for both the encoder and the decoder.”
Improving Neural Question Generation using Answer Separation (2018, methodology): “We also ﬁgure out an interesting observation: even though (Song et al. 2018) is the previous state-of-the-art NQG model, it generates more irrelevant questions to the target answer when compared to seq2seq+AP.”
A framework for automatic question generation from text using deep reinforcement learning (2018, methodology): “(Ranzato et al., 2015) incrementally trains for the first s steps on the cross entropy loss then for the subsequent t-s steps, it trains using REINFORCE (Sutton and Barto, 1998) and this process repeats.”