PreGNN: Hardware Acceleration to Take Preprocessing Off the Critical Path in Graph Neural Networks.
Donghyun GoukSeungkwan KangMiryeong KwonJunhyeok JangHyunkyu ChoiSangwon LeeMyoungsoo JungPublished in: IEEE Comput. Archit. Lett. (2022)
Keyphrases
- critical path
- preprocessing
- neural network
- job shop scheduling problem
- low cost
- graph model
- preprocessing stage
- graph theory
- hardware and software
- pattern recognition
- fuzzy logic
- artificial neural networks
- genetic algorithm
- image processing
- graph representation
- real time
- random walk
- back propagation
- graph structure
- hardware implementation
- computer systems
- scheduling problem
- memetic algorithm