Question Answering by Reasoning Across Documents with Graph Convolutional Networks

By Nicola De Cao, Wilker Aziz, Ivan Titov. 2018

245 citations recorded by the source corpus.

Most research in reading comprehension has focused on answering questions based on individual documents or even single paragraphs. We introduce a neural model which integrates and reasons relying on information spread within documents and across multiple documents. We frame it as an inference problem on a graph. Mentions of entities are nodes of this graph while edges encode relations between different mentions (e.g., within- and cross-document co-reference). Graph convolutional networks (GCNs) are applied to these graphs and trained to perform multi-step reasoning. Our Entity-GCN method is scalable and compact, and it achieves state-of-the-art results on a multi-document question answering dataset, WikiHop (Welbl et al., 2018).

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Natural Language Processing, Graph 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 Computer Vision and Pattern Recognition

In Natural Language Processing

In Graph Learning

Citations in the corpus

Knowledge Guided Text Retrieval and Reading for Open Domain Question Answering (2019, background): “The most related work to ours includes Song et al. (2018) and Cao et al. (2019), which construct a graph of entities in the passage through entity detection and coreference resolution.”
Multi-hop Reading Comprehension through Question Decomposition and Rescoring (2019, methodology): “Prior methods on multi-hop reading comprehension focus on answering relational queries, and emphasize attention models that reason over coreference chains (Dhingra et al., 2018; Zhong et al., 2019; Cao et al., 2019).”
Answering Complex Open-Domain Questions with Multi-Hop Dense Retrieval (2020, background): “…either retrieve passages from a small passage pool pruned based on the the speciﬁc dataset (Sun et al., 2019; Dhingra et al., 2020), or focus on a non-retrieval setting where a compact documents set is already given (De Cao et al., 2018; Zhong et al., 2019; Tu et al., 2019; Beltagy et al., 2020).”
Hierarchical Graph Network for Multi-hop Question Answering (2019, background): “…of the dataset, answers can be selected either from the entities in the constructed entity graph (Song et al., 2018; Dhin-gra et al., 2018; De Cao et al., 2019; Tu et al., 2019; Ding et al., 2019), or from spans of documents by fusing entity representations back into token-level document…”
Multi-hop Question Answering via Reasoning Chains (2019, methodology): “More recent work including Entity-GCN (De Cao et al., 2018), MHQA-GRN (Song et al., 2018), and BAG (Cao et al., 2019), form this problem as a search over entity graph, and adapt graph convolution network (Kipf and Welling, 2017) to do reasoning.”
Structured Neural Summarization (2018, background): “De Cao et al. (2018) create a graphover named entities over a set of documents to assist question answering.”
BertGCN: Transductive Text Classification by Combining GNN and BERT (2021, background): “…of end tasks such as relation extraction (Zhang et al., 2018b), semantic role labeling (Marcheggiani and Titov, 2017), data-to-text generation (Marcheggiani and Perez-Beltrachini, 2018), machine translation (Bastings et al., 2017) and question answering (Song et al., 2018; De Cao et al., 2018).”
Do Multi-hop Readers Dream of Reasoning Chains? (2019, unknown): “More recent development of QA systems (Song et al., 2018; De Cao et al., 2018; Zhong et al., 2019) has started to focus on multi-hop reasoning on text passages, aiming to propose more sophisticated models beyond the shallow matching between questions and answers.”
Reasoning with Latent Structure Refinement for Document-Level Relation Extraction (2020, methodology): “We adapt rules from De Cao et al. (2019) for multi-hop question answering, i.e., each mention node is connected to its entity node and to the same mention nodes across sentences, while mention nodes and MDP nodes which reside in the same sentence are fully connected.”
Global Reasoning over Database Structures for Text-to-SQL Parsing (2019, methodology): “A graph convolutional network (GCN) then learned representations h v for nodes end-to-end (De Cao et al., 2019; Sorokin and Gurevych, 2018).”
Improving Robustness by Augmenting Training Sentences with Predicate-Argument Structures (2020, background): “Similarly, Cao et al. (2019) and Dhingra et al. (2018) incorporate linguistic information, i.e., coreference relations, in their model and show improvements in in-domain evaluations.”
Relational Representation Learning for Dynamic (Knowledge) Graphs: A Survey (2019, background): “De Cao et al. [56] considered question answering by reasoning over multiple documents.”
Rethinking Search: Making Experts out of Dilettantes (2021, background): “Within the context of neural retrieval from an encoder-decoder model, it may also be possible to incorporate relational inductive biases [2, 18] to model relationships amongst candidate documents and terms in the decoder.”