mlech26l/ncps
A PyTorch and TensorFlow library implementing Neural Circuit Policies (NCP), Liquid Time-Constant (LTC), and Closed-form Continuous-time (CfC) recurrent neural networks inspired by C. elegans nervous system.
Not currently ranked — collecting fresh signals.
star history
Neural Circuit Policies (NCPs) are sparse recurrent neural networks designed for auditable autonomy and efficient time-series processing. The library provides reference implementations of NCP, LTC, and CfC architectures in both PyTorch and TensorFlow/Keras, including utilities for working with irregularly sampled time-series data. Published work includes papers in Nature Machine Intelligence on closed-form continuous-time neural networks.
Frequently asked
- What is mlech26l/ncps?
- A PyTorch and TensorFlow library implementing Neural Circuit Policies (NCP), Liquid Time-Constant (LTC), and Closed-form Continuous-time (CfC) recurrent neural networks inspired by C. elegans nervous system.
- Is ncps open source?
- Yes — mlech26l/ncps is open source, released under the Apache-2.0 license.
- What language is ncps written in?
- mlech26l/ncps is primarily written in Python.
- How popular is ncps?
- mlech26l/ncps has 2.3k stars on GitHub.
- Where can I find ncps?
- mlech26l/ncps is on GitHub at https://github.com/mlech26l/ncps.