Prediction of Inference Energy on CNN Accelerators Supporting Approximate Circuits.
Michal PinosVojtech MrazekLukás SekaninaPublished in: DDECS (2023)
Keyphrases
- prediction accuracy
- cellular neural networks
- energy consumption
- real time
- prediction error
- prediction model
- probabilistic inference
- bayesian inference
- analog circuits
- computing systems
- prediction algorithm
- inference process
- inference in graphical models
- convolutional neural network
- single chip
- energy minimization
- main memory
- motion estimation