CoGNN: An Algorithm-Hardware Co-Design Approach to Accelerate GNN Inference With Minibatch Sampling.
Kai ZhongShulin ZengWentao HouGuohao DaiZhenhua ZhuXuecang ZhangShihai XiaoHuazhong YangYu WangPublished in: IEEE Trans. Comput. Aided Des. Integr. Circuits Syst. (2023)
Keyphrases
- experimental evaluation
- dynamic programming
- hardware implementation
- worst case
- detection algorithm
- high accuracy
- theoretical analysis
- learning algorithm
- optimal solution
- computational cost
- uniform sampling
- inference process
- convergence rate
- times faster
- monte carlo
- segmentation algorithm
- linear programming
- low cost
- probabilistic model
- significant improvement
- k means
- objective function
- optimization algorithm
- parameter space
- hardware and software
- markov chain monte carlo
- sampling algorithm
- decision trees
- real time