Gibbs Sampling with Low-Power Spiking Digital Neurons.
Srinjoy DasBruno U. PedroniPaul MerollaJohn V. ArthurAndrew S. CassidyBryan L. JacksonDharmendra S. ModhaGert CauwenberghsKenneth Kreutz-DelgadoPublished in: CoRR (2015)
Keyphrases
- low power
- gibbs sampling
- mixed signal
- neuron model
- high speed
- power consumption
- low cost
- spike trains
- hodgkin huxley
- spiking neurons
- markov chain
- spiking neural networks
- single neuron
- parameter estimation
- topic models
- approximate inference
- em algorithm
- latent dirichlet allocation
- expectation maximization
- low power consumption
- neural network
- belief networks
- markov chain monte carlo
- exact inference
- logic circuits
- image sensor
- cmos technology
- multi channel
- graphical models
- ultra low power
- higher order
- artificial neural networks
- probabilistic inference
- latent variables
- belief propagation
- maximum likelihood
- state space
- bayesian networks