Description
Ordinary RNNs for time series and control are parameter-heavy and hard to interpret. Neural Circuit Policies provides biologically inspired NCP, LTC and CfC neural models, published in Nature Machine Intelligence.
It supports PyTorch and TensorFlow, achieving robust control and sequence modeling with very few neurons.
Bio-inspired:Neural circuits.
Liquid nets:LTC and CfC.
Compact:Interpretable.
Frameworks:PyTorch and TensorFlow.
It supports PyTorch and TensorFlow, achieving robust control and sequence modeling with very few neurons.
Features
Bio-inspired:Neural circuits.
Liquid nets:LTC and CfC.
Compact:Interpretable.
Frameworks:PyTorch and TensorFlow.
