Combining evolution and self-organization to find natural Boolean representations in unconventional computational...

Area of Science:

• Computational intelligence
• Artificial computing systems
• Bio-inspired computing

Background:

• Designing unconventional computing systems requires optimizing both computational structure and symbol encoding.
• Current methods using heuristic search and evolutionary algorithms are time-consuming due to interdependencies.
• A more efficient approach is needed for developing novel computing paradigms.

Purpose of the Study:

• To introduce a novel approach combining evolution with self-organization for designing unconventional computing systems.
• To enable the emergence of symbol encoding through self-organization while evolving computational structures.
• To demonstrate the efficiency and effectiveness of this combined approach.

Main Methods:

• Evolving computational structures using an evolutionary algorithm.
• Allowing symbol encoding (representations for TRUE and FALSE) to emerge via self-organization.
• Implementing and testing the approach on a continuous-time recurrent neural network (CTRNN) and BZ-droplet-based computing (DropSim) models.

Main Results:

• Successfully demonstrated NAND-gate functionality in both CTRNN and DropSim models.
• Identified natural representations for TRUE and FALSE symbols through self-organization.
• Validated the performance of evolved NAND gates in half-adder networks, showing correct outputs for all input combinations and transitions.

Conclusions:

• Combining evolution with self-organization is an efficient strategy for designing unconventional computing systems.
• This approach leads to more natural symbol encodings.
• The demonstrated technique holds broad applicability beyond the specific systems tested.