Liquid-Resistance Liquid-Capacitance (LRC) Networks

This repository contains code for the paper "Liquid-Resistance Liquid-Capacitance Networks" presented at the NeuroAI Workshop at NeurIPS 2024. The paper is available on ArXiv.

Classification

To run the classification examples:

cd classification
python run_imdb.py --model LRC_sym_elastance --size 64

Model choices are LRCs with two types of elastance: LRC_sym_elastance and LRC_asym_elastance, and lstm, mgu, gru.

For the person localization, first download the dataset by running the download_dataset.sh script.

Neural ODE

To run the neural ODE examples:

cd neuralODE
python run_ode.py --model lrc --lrc_type symmetric --data spiral --niters 1000

The data choices are: periodic_sinusodial, spiral, duffing, periodic_predator_prey, limited_predator_prey, nonlinear_predator_prey.

Use --viz True for visualizing the progress of each validation step.

Sinusoid	Spiral	Duffing

Periodic Lotka-Volterra	Limited Lotka-Volterra	Non-linear Lotka-Volterra

Citation

If you use this work, please cite our paper as follows:

@misc{farsang2024liquidresistanceliquidcapacitance,
      title={Liquid Resistance Liquid Capacitance Networks}, 
      author={Mónika Farsang and Sophie A. Neubauer and Radu Grosu},
      year={2024},
      eprint={2403.08791},
      archivePrefix={arXiv},
      primaryClass={cs.NE},
      url={https://arxiv.org/abs/2403.08791}, 
}