Artificial Epigenetic Networks: Automatic Decomposition of Dynamical Control Tasks Using Topological...

Area of Science:

• Computational neuroscience
• Artificial intelligence
• Systems biology

Background:

• Inspired by gene regulatory networks and chromatin remodeling in biological organisms.
• Addresses the challenge of solving dynamical problems without prior structural information.

Purpose of the Study:

• To introduce and evaluate the artificial epigenetic network (AEN).
• To demonstrate AEN's ability to automatically decompose and solve dynamical problems through topological self-modification.

Main Methods:

• Developed a recurrent connectionist architecture (AEN) capable of dynamic topology modification.
• Tested the AEN on a series of dynamical control tasks requiring adaptive behavioral switching.

Main Results:

• AENs significantly improved performance and robustness compared to fixed-topology controllers.
• Topological changes were shown to stabilize attractors, facilitating stable dynamical regimes and rapid switching.
• Network partitioning provided novel insights into problem structures.

Conclusions:

• Artificial epigenetic networks offer a robust and sensitive approach for adaptive control in complex dynamical systems.
• The self-modifying topology is key to stabilizing and switching between different dynamical behaviors.
• This architecture holds potential for applications requiring adaptable problem-solving capabilities.