SoftSNN: low-cost fault tolerance for spiking neural network accelerators under soft errors.
Rachmad Vidya Wicaksana PutraMuhammad Abdullah HanifMuhammad ShafiquePublished in: DAC (2022)
Keyphrases
- fault tolerance
- low cost
- spiking neural networks
- fault tolerant
- error detection
- single chip
- biologically inspired
- distributed systems
- load balancing
- cerebellar model
- distributed computing
- response time
- group communication
- mobile agents
- peer to peer
- liquid state machine
- replicated databases
- feed forward
- spiking neurons
- database replication
- neuron model
- biologically plausible
- fault management
- computing systems
- motor control
- high performance computing
- real time
- multi modal
- single point of failure
- artificial neural networks
- machine learning