Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2402.18875 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917662212227072 |
|---|---|
| author | Wong, Zhen Hao Yang, Hansi Fu, Xiaoyi Yao, Quanming |
| author_facet | Wong, Zhen Hao Yang, Hansi Fu, Xiaoyi Yao, Quanming |
| contents | Heterogeneous Graph Neural Networks (HGNNs) are a class of deep learning models designed specifically for heterogeneous graphs, which are graphs that contain different types of nodes and edges. This paper investigates the application of curriculum learning techniques to improve the performance and robustness of Heterogeneous Graph Neural Networks (GNNs). To better classify the quality of the data, we design a loss-aware training schedule, named LTS that measures the quality of every nodes of the data and incorporate the training dataset into the model in a progressive manner that increases difficulty step by step. LTS can be seamlessly integrated into various frameworks, effectively reducing bias and variance, mitigating the impact of noisy data, and enhancing overall accuracy. Our findings demonstrate the efficacy of curriculum learning in enhancing HGNNs capabilities for analyzing complex graph-structured data. The code is public at https://github.com/LARS-research/CLGNN/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_18875 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Loss-aware Curriculum Learning for Heterogeneous Graph Neural Networks Wong, Zhen Hao Yang, Hansi Fu, Xiaoyi Yao, Quanming Machine Learning Heterogeneous Graph Neural Networks (HGNNs) are a class of deep learning models designed specifically for heterogeneous graphs, which are graphs that contain different types of nodes and edges. This paper investigates the application of curriculum learning techniques to improve the performance and robustness of Heterogeneous Graph Neural Networks (GNNs). To better classify the quality of the data, we design a loss-aware training schedule, named LTS that measures the quality of every nodes of the data and incorporate the training dataset into the model in a progressive manner that increases difficulty step by step. LTS can be seamlessly integrated into various frameworks, effectively reducing bias and variance, mitigating the impact of noisy data, and enhancing overall accuracy. Our findings demonstrate the efficacy of curriculum learning in enhancing HGNNs capabilities for analyzing complex graph-structured data. The code is public at https://github.com/LARS-research/CLGNN/. |
| title | Loss-aware Curriculum Learning for Heterogeneous Graph Neural Networks |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2402.18875 |