We are presenting two papers at the International Conference on Very Large Databses (VLDB) 2024.

In the first work, Everest: GPU-Accelerated System For Mining Temporal Motifs, we designed GPU-based system to mine temporal motifs. We characterized the temporal motif mining workload on the GPU, and designed hardware-specific optimizations. This work was led by the insanely-talented Yichao Yuan and was done in collaboration with colleagues from University of Michigan.

In the second work, Demystifying Graph Sparsification Algorithms in Graph Properties Preservation, we ask the question: what is the best way to sparsify the graph? The short answer: no one size fits all. This is the culmination of more than two years of effort, led by the relentless Yuhan Chen, and it was done in collaboration with colleagues from Technion and University of Michigan.