Welcome to Neural Circuit Policies’s documentation!

Neural Circuit Policies (NCPs) are machine learning models inspired by the nervous system of the nematode C. elegans. This package provides easy-to-use implementations of NCPs for PyTorch and Tensorflow.

pip3 install -U ncps

Example Pytorch example:

from ncps.torch import CfC

# a fully connected CfC network
rnn = CfC(input_size=20, units=50)
x = torch.randn(2, 3, 20) # (batch, time, features)
h0 = torch.zeros(2,50) # (batch, units)
output, hn = rnn(x,h0)

A Tensorflow example

# Tensorflow example
from ncps.tf import LTC
from ncps.wirings import AutoNCP

wiring = AutoNCP(28, 4) # 28 neurons, 4 outputs
model = tf.keras.models.Sequential(
    [
        tf.keras.layers.InputLayer(input_shape=(None, 2)),
        # LTC model with NCP sparse wiring
        LTC(wiring, return_sequences=True),
    ]
)

User’s Guide

• Quickstart
  • Neuron Models
  • Wirings
  • Diagram
• Examples
  • First steps (Pytorch)
  • First steps (Tensorflow)
  • Partially Observabily Reinforcement Learning
  • Atari Behavior Cloning
  • Atari Reinforcement Learning (PPO)
• API Reference
  • PyTorch (nn.modules)
  • Tensorflow (tf.keras Layers)
  • Wirings