Current browse context:
cs
Change to browse by:
References & Citations
Computer Science > Computation and Language
Title: Generic and Trend-aware Curriculum Learning for Relation Extraction in Graph Neural Networks
(Submitted on 17 May 2022 (v1), last revised 8 Jun 2022 (this version, v2))
Abstract: 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.
Submission history
From: Nidhi Vakil [view email][v1] Tue, 17 May 2022 20:46:02 GMT (2823kb,D)
[v2] Wed, 8 Jun 2022 14:34:26 GMT (2860kb,D)
Link back to: arXiv, form interface, contact.