Abstract:The task of machine reading comprehension (MRC) is to enable machine to read and understand a piece of text, and then answer the corresponding question correctly. This task requires machine to not only be able to perform semantic understanding, but also possess logical reasoning capabilities. Just like human reading, it involves thinking about the text from two interacting perspectives of semantics and logic. However, previous methods based on reading comprehension either consider only the logical structure of the text or only the semantic structure of the text, and cannot simultaneously balance semantic understanding and logical reasoning. This single form of reasoning cannot make the machine fully understand the meaning of the text. Additionally, the issue of sparsity in composition presents a significant challenge for models that rely on graph-based reasoning. To this end, a cross-graph knowledge propagation network (CGKPN) with adaptive connection is presented to address the above issues. The model first performs self-view node embedding on the constructed logical graph and semantic graph to update the representations of the graphs. Specifically, relevance matrix between nodes is introduced to adaptively adjust node connections in response to the challenge posed by sparse graph. Subsequently, CGKPN conducts cross-graph knowledge propagation on nodes that are identical in both graphs, effectively resolving conflicts arising from identical nodes in different views, and enabling the model to better integrate the logical and semantic relationships of the text through efficient interaction. Experiments on the two MRC datasets ReClor and LogiQA indicate the superior performance of our proposed model CGKPN compared to other existing baselines.

M-HGN: Multi-information Enhanced Heterogeneous Graph Network for Multi-party Dialogue Reading Comprehension

Dialogue Logic Aware and Key Utterance Decoupling Model for Multi-Party Dialogue Reading Comprehension

GSN: A Graph-Structured Network for Multi-Party Dialogues

Multi-hop Reading Comprehension across Multiple Documents by Reasoning over Heterogeneous Graphs

Dialogue Relation Extraction with Document-Level Heterogeneous Graph Attention Networks

Discovering Multimodal Hierarchical Structures with Graph Neural Networks for Multi-modal and Multi-hop Question Answering.

Multi-Matching Network for Multiple Choice Reading Comprehension

Multi-modal multi-hop interaction network for dialogue response generation

Incorporating Commonsense Knowledge into Abstractive Dialogue Summarization via Heterogeneous Graph Networks

HETERMPC: A Heterogeneous Graph Neural Network for Response Generation in Multi-Party Conversations

Heterogeneous Relational Graph Neural Networks with Adaptive Objective for End-to-End Task-Oriented Dialogue

Multi-Modal Graph Neural Network for Joint Reasoning on Vision and Scene Text

ClueReader: Heterogeneous Graph Attention Network for Multi-hop Machine Reading Comprehension

MHGNN: Multi-view fusion based Heterogeneous Graph Neural Network

Dialogue Graph Modeling for Conversational Machine Reading

Hierarchical Graph Network for Multi-hop Question Answering

Multi-turn dialogue comprehension from a topic-aware perspective

CGKPN: Cross-Graph Knowledge Propagation Network with Adaptive Connection for Reasoning-Based Machine Reading Comprehension

Seeing both sides: context-aware heterogeneous graph matching networks for extracting-related arguments

Multi-hop Reading Comprehension Across Documents with Path-based Graph Convolutional Network

GCNet: Graph Completion Network for Incomplete Multimodal Learning in Conversation