HiPool: Modeling Long Documents Using Graph Neural Networks
This addresses the challenge of processing long documents for NLP applications, offering a scalable solution that outperforms existing hierarchical sequential models, though it is incremental in advancing graph-based approaches for this domain.
The paper tackles the problem of encoding long sequences in NLP by proposing HiPool, a graph-based method that models intra- and cross-sentence correlations to overcome long dependency issues in hierarchical models, achieving a 2.6% improvement in F1 score over baselines and 4.8% on the longest sequence dataset.
Encoding long sequences in Natural Language Processing (NLP) is a challenging problem. Though recent pretraining language models achieve satisfying performances in many NLP tasks, they are still restricted by a pre-defined maximum length, making them challenging to be extended to longer sequences. So some recent works utilize hierarchies to model long sequences. However, most of them apply sequential models for upper hierarchies, suffering from long dependency issues. In this paper, we alleviate these issues through a graph-based method. We first chunk the sequence with a fixed length to model the sentence-level information. We then leverage graphs to model intra- and cross-sentence correlations with a new attention mechanism. Additionally, due to limited standard benchmarks for long document classification (LDC), we propose a new challenging benchmark, totaling six datasets with up to 53k samples and 4034 average tokens' length. Evaluation shows our model surpasses competitive baselines by 2.6% in F1 score, and 4.8% on the longest sequence dataset. Our method is shown to outperform hierarchical sequential models with better performance and scalability, especially for longer sequences.