Skip to content

Latest commit

 

History

History
69 lines (45 loc) · 3.38 KB

README.md

File metadata and controls

69 lines (45 loc) · 3.38 KB
Raw

dynamic-neural-field-degeneration

Alt text

Overview

dynamic-neural-field-degeneration is a framework designed to simulate and study the effects of artificial neural degeneration on Dynamic Neural Field (DNF) models. This work explores the resilience and adaptability of DNFs, specifically in the context of cognitive behavior in robotic systems. The project centers on a robotic sorting task where degeneration's impact on task performance is tested and visualized.

Key objectives:

  • Investigate resilience of DNFs under neuron death and synaptic disruption.
  • Explore adaptability through relearning mechanisms to recover functionality post-degeneration.
  • Assess biological analogues by modeling degenerative effects inspired by conditions such as Huntington’s, Alzheimer’s, and Parkinson’s diseases.

Features

  • Dynamic neural field simulation and visualization: This project runs dynamic-neural-field-composer for the creation composition, and real-time visualization of neural field dynamics.
  • Degeneration modeling: Induce artificial neurodegenerative conditions (e.g., neuron death, synaptic connection degradation) and observe their impact.
  • Relearning and recovery: Implement mechanisms to reorganize connections and recover functionality in degraded architectures.
  • Robotic sorting task integration: The experiment is grounded in a simulated, using CoppeliaSim, sorting task with a UR10 robotic manipulator.

Installation

To run this project, ensure you have the following dependencies installed:

  1. dynamic-neural-field-composer: Clone the repository and follow the provided building and installation instructions.
  2. coppeliasim-cpp: Again, clone the repository and follow its building and installation instructions.

Once dependencies are installed, you can build this project using the build.bat file provided in the root directory.

Executable Files:

  • To simulate degeneration, run inducing-degeneration.exe.
  • To simulate recovery and relearning, run recovering-from-degeneration.exe.

Usage

To replicate specific experiments, switch to the corresponding Git branch, as follows:

  • Inducing Degeneration: git checkout inducing-degeneration.
  • Relearning from Degeneration: git checkout recovering-from-degeneration.

Modifying Experiment Parameters

To adjust the simulation parameters for either experiment, edit the experiment_parameters.json file to configure factors such as neuron behavior, synaptic degradation rates, or other simulation settings.

Viewing the Robotic Simulation

For the relearning experiment, which is coupled with a robotic simulation, you can view the sorting task:

  1. git checkout coppeliasim-visualization.
  2. Open the scenario.ttt scene file in CoppeliaSim.
  3. Press Play in CoppeliaSim before or after running the recovering-from-degeneration.exe executable to observe the robotic system performing the task.

Video explanation

Watch the video

Documentation

For a full exploration of the repository refer to the Wiki.