<ul><li>Graph convolutional networks (GCNs) are popular for building machine-learning applications for graph-structured data.</li><li>This work introduces GCN-ABFT, a cost-effective approach for error detection in GCN accelerators.</li><li>GCN-ABFT calculates a checksum for the entire three-matrix product in a single GCN layer.</li><li>Experimental results show that GCN-ABFT reduces the number of operations needed for checksum computation while maintaining fault-detection accuracy.</li></ul>

GCN-ABFT: Low-Cost Online Error Checking for Graph Convolutional Networks

Discover more