Message Passing Attention Networks for Document Understanding
Authors:
- Antoine J. -P. TixierATAntoine J. -P. Tixier
Preprint
Abstract
Most graph neural networks can be described in terms of message passing, vertex update, and readout functions. In this paper, we represent documents as word co-occurrence networks and propose an application of the message passing framework to NLP, the Message Passing Attention network for Document understanding (MPAD). We also propose several hierarchical variants of MPAD. Experiments conducted on 10 standard text classification datasets show that our architectures are competitive with the state-of-the-art. Ablation studies reveal further insights about the impact of the different components on performance. Code and data are publicly available.Keywords
All Related Versions
This publication has 7 references indexed in Scilit:
- Energy-based Self-attentive Learning of Abstractive Communities for Spoken Language Understanding2020
- Graph Convolutional Networks for Text ClassificationProceedings of the AAAI Conference on Artificial Intelligence, 2019
- Dialogue Act Classification with Context-Aware Self-Attention2019
- Molecular graph convolutions: moving beyond fingerprintsJournal of Computer-Aided Molecular Design, 2016
- A Hierarchical Neural Autoencoder for Paragraphs and Documents2015
- Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation2014
- Annotating Expressions of Opinions and Emotions in LanguageLanguage Resources and Evaluation, 2005