A 5.37-TSOPS/W Reconfigurable Neuron Array with Dual-mode Neurons and Asynchronous Synapses for Energy-Efficient Inference and Biological Neural Network Simulation.
Xiangao QiYuqing LouYongfu LiGuoxing WangKea-Tiong TangJian ZhaoPublished in: A-SSCC (2023)
Keyphrases
- efficient inference
- neural network
- spiking neurons
- biologically plausible
- spiking neural networks
- synaptic weights
- neuron model
- biologically inspired
- synaptic plasticity
- associative memory
- feed forward
- neural models
- output layer
- probabilistic inference
- artificial neural networks
- fully connected
- activation function
- hidden layer
- learning rules
- hebbian learning
- biological systems
- markov random field
- back propagation
- spike trains
- conditional random fields
- neural network model
- network architecture
- neural model
- visual cortex
- human pose estimation
- hidden variables
- support vector
- input patterns
- factor graphs
- neural nets
- recurrent neural networks
- receptive fields
- junction tree
- approximate inference
- markov networks