Robust Training of Temporal GNNs using Nearest Neighbours based Hard Negatives
arXiv:2402.09239 · doi:10.1145/3632410.3632464
Abstract
Temporal graph neural networks Tgnn have exhibited state-of-art performance in future-link prediction tasks. Training of these TGNNs is enumerated by uniform random sampling based unsupervised loss. During training, in the context of a positive example, the loss is computed over uninformative negatives, which introduces redundancy and sub-optimal performance. In this paper, we propose modified unsupervised learning of Tgnn, by replacing the uniform negative sampling with importance-based negative sampling. We theoretically motivate and define the dynamically computed distribution for a sampling of negative examples. Finally, using empirical evaluations over three real-world datasets, we show that Tgnn trained using loss based on proposed negative sampling provides consistent superior performance.
arXiv comment10 pages
References in corpus (7)
- Graph Convolutional Neural Networks for Web-Scale Recommender Systems
- Predicting Dynamic Embedding Trajectory in Temporal Interaction Networks
- Community Interaction and Conflict on the Web
- Node Embedding over Temporal Graphs
- Contrastive Learning with Adversarial Examples
- Optimizing Dense Retrieval Model Training with Hard Negatives
- Privacy-Aware Recommender Systems Challenge on Twitter's Home Timeline