How Can Graph Neural Networks Help Document Retrieval: A Case Study on CORD19 with Concept Map Generation

  title={How Can Graph Neural Networks Help Document Retrieval: A Case Study on CORD19 with Concept Map Generation},
  author={Hejie Cui and Jiaying Lu and Yao Ge and Carl Yang},
Graph neural networks (GNNs), as a group of powerful tools for representation learning on irregular data, have manifested superiority in various downstream tasks. With unstructured texts represented as concept maps, GNNs can be exploited for tasks like document retrieval. Intrigued by how can GNNs help document retrieval, we conduct an empirical study on a large-scale multi-discipline dataset CORD19. Results show that instead of the complex structure-oriented GNNs such as GINs and GATs, our… 


On Positional and Structural Node Features for Graph Neural Networks on Non-attributed Graphs
The two types of artificial node features are pointed out, i.e., positional and structural node features, and insights on why each of them is more appropriate for certain tasks are provided, leading to a practical guideline on the choices between different artificial nodes features for GNNs on non-attributed graphs.
Neural Concept Map Generation for Effective Document Classification with Interpretable Structured Summarization
Concept maps provide concise structured representations for documents regarding their important concepts and interaction links, which have been widely used for document summarization and downstream
How Powerful are Graph Neural Networks?
This work characterize the discriminative power of popular GNN variants, such as Graph Convolutional Networks and GraphSAGE, and show that they cannot learn to distinguish certain simple graph structures, and develops a simple architecture that is provably the most expressive among the class of GNNs.
Graph Convolutional Neural Networks for Web-Scale Recommender Systems
A novel method based on highly efficient random walks to structure the convolutions and a novel training strategy that relies on harder-and-harder training examples to improve robustness and convergence of the model are developed.
IR-BERT: Leveraging BERT for Semantic Search in Background Linking for News Articles
The empirically show that employing a language model benefits the approach in understanding the context as well as the background of the query article, and a diversity measure is proposed to evaluate the effectiveness of the various approaches in retrieving a diverse set of documents.
Graph Attention Networks
We present graph attention networks (GATs), novel neural network architectures that operate on graph-structured data, leveraging masked self-attentional layers to address the shortcomings of prior
A Graph-based Relevance Matching Model for Ad-hoc Retrieval
This work proposes a novel relevance matching model based on graph neural networks to leverage the documentlevel word relationships for ad-hoc retrieval and explicitly incorporate all contexts of a term through the graph-of-word text format.
Graph based model for information retrieval using a stochastic local search
Connecting the Dots: Event Graph Schema Induction with Path Language Modeling
This work proposes a new Event Graph Schema, where two event types are connected through multiple paths involving entities that fill important roles in a coherent story, and introduces Path Language Model, an auto-regressive language model trained on event-event paths, to select salient and coherent paths to probabilistically construct these graph schemas.
MultiSage: Empowering GCN with Contextualized Multi-Embeddings on Web-Scale Multipartite Networks
A contextualized GCN engine is presented by modeling the multipartite networks of target nodes and their intermediatecontext nodes that specify the contexts of their interactions to achieve interaction contextualization by treating neighboring target nodes based on intermediate context nodes.