Message Passing Attention Networks for Document Understanding - Archive ouverte HAL
Communication Dans Un Congrès Année : 2020

Message Passing Attention Networks for Document Understanding

Résumé

Graph neural networks have recently emerged as a very effective framework for processing graph-structured data. These models have achieved state-of-the-art performance in many tasks. 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 is publicly available at: https://github.com/giannisnik/mpad.

Dates et versions

hal-04492856 , version 1 (06-03-2024)

Identifiants

Citer

Giannis Nikolentzos, Antoine Tixier, Michalis Vazirgiannis. Message Passing Attention Networks for Document Understanding. The Thirty-Fourth AAAI Conference on Artificial Intelligence, Feb 2020, New York, NY, United States. pp.8544-8551, ⟨10.1609/aaai.v34i05.6376⟩. ⟨hal-04492856⟩
12 Consultations
0 Téléchargements

Altmetric

Partager

More