Improving DNN Fault Tolerance using Weight Pruning and Differential Crossbar Mapping for ReRAM-based Edge AI.
Geng YuanZhiheng LiaoXiaolong MaYuxuan CaiZhenglun KongXuan ShenJingyan FuZhengang LiChengming ZhangHongwu PengNing LiuAo RenJinhui WangYanzhi WangPublished in: CoRR (2021)
Keyphrases
- fault tolerance
- fault tolerant
- load balancing
- distributed systems
- high availability
- distributed computing
- artificial intelligence
- response time
- group communication
- peer to peer
- search space
- expert systems
- mobile agents
- replicated databases
- database replication
- high scalability
- fault management
- single point of failure
- data replication
- scheduling algorithm
- edge detection
- knowledge representation
- error detection
- data sets
- high performance computing
- weighted graph
- failure recovery