MUSTARD: a coupled, stochastic/deterministic, discrete/continuous technique for predicting the impact of random telegraph noise on SRAMs and DRAMs.
Aadithya V. KarthikSriramkumar VenugopalanAlper DemirJaijeet S. RoychowdhuryPublished in: DAC (2011)
Keyphrases
- continuous domains
- sampled data
- discrete space
- continuous state spaces
- continuous variables
- discrete variables
- continuous action
- stochastic optimization problems
- continuous space
- stochastic methods
- markov processes
- noisy data
- missing data
- discrete data
- noise level
- random noise
- signal to noise ratio
- continuous functions
- continuous data
- noise model
- additive noise
- discrete geometry
- continuous valued
- monte carlo
- black box
- stochastic domains
- stochastic optimization
- discrete version
- reinforcement learning
- discrete valued
- noise sensitivity
- piecewise linear
- gaussian noise
- learning automata