Generic and Trend-aware Curriculum Learning for Relation Extraction in Graph Neural Networks
This work addresses relation extraction in text graphs, which is important for natural language processing applications, but it appears incremental as it extends existing curriculum learning approaches.
The authors tackled the problem of relation extraction in graph neural networks by developing a curriculum learning approach that uses sample-level loss trends to schedule training samples, achieving sizable improvements over state-of-the-art methods across multiple datasets.
We present a generic and trend-aware curriculum learning approach for graph neural networks. It extends existing approaches by incorporating sample-level loss trends to better discriminate easier from harder samples and schedule them for training. The model effectively integrates textual and structural information for relation extraction in text graphs. Experimental results show that the model provides robust estimations of sample difficulty and shows sizable improvement over the state-of-the-art approaches across several datasets.