GraSAME: Injecting Token-Level Structural Information to Pretrained Language Models via Graph-guided Self-Attention Mechanism
arxiv(2024)
摘要
Pretrained Language Models (PLMs) benefit from external knowledge stored in
graph structures for various downstream tasks. However, bridging the modality
gap between graph structures and text remains a significant challenge.
Traditional methods like linearizing graphs for PLMs lose vital graph
connectivity, whereas Graph Neural Networks (GNNs) require cumbersome processes
for integration into PLMs. In this work, we propose a novel graph-guided
self-attention mechanism, GraSAME. GraSAME seamlessly incorporates token-level
structural information into PLMs without necessitating additional alignment or
concatenation efforts. As an end-to-end, lightweight multimodal module, GraSAME
follows a multi-task learning strategy and effectively bridges the gap between
graph and textual modalities, facilitating dynamic interactions between GNNs
and PLMs. Our experiments on the graph-to-text generation task demonstrate that
GraSAME outperforms baseline models and achieves results comparable to
state-of-the-art (SOTA) models on WebNLG datasets. Furthermore, compared to
SOTA models, GraSAME eliminates the need for extra pre-training tasks to adjust
graph inputs and reduces the number of trainable parameters by over 100
million.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要