A Heuristic and Greedy Weight Remapping Scheme with Hardware Optimization for Irregular Sparse Neural Networks Implemented on CIM Accelerator in Edge AI Applications.
Lizhou WuChenyang ZhaoJingbo WangXueru YuShoumian ChenChen LiJun HanXiaoyong XueXiaoyang ZengPublished in: ASPDAC (2024)
Keyphrases
- neural network
- search algorithm
- greedy algorithm
- optimization scheme
- real time
- dynamic programming
- hardware and software
- optimization problems
- optimization method
- protection scheme
- combinatorial optimization
- fuzzy logic
- genetic algorithm
- squeaky wheel
- hill climbing
- field programmable gate array
- greedy heuristic
- stochastic search
- optimization algorithm
- greedy algorithms
- back propagation
- sparse representation
- locally optimal
- optimal solution
- vlsi implementation
- weight update
- highly non linear
- greedy search
- feature selection
- weighted graph
- parallel implementation
- embedded systems
- multilayer perceptron
- decision trees
- edge detection
- edge information
- edge weights
- multi objective
- image processing
- pattern recognition
- computer systems
- artificial neural networks
- low cost
- evolutionary algorithm
- scheduling problem