Abstract:Biological plastic neural networks are systems of extraordinary computational capabilities shaped by evolution, development, and lifetime learning. The interplay of these elements leads to the emergence of adaptive behavior and intelligence. Inspired by such intricate natural phenomena, Evolved Plastic Artificial Neural Networks (EPANNs) use simulated evolution in-silico to breed plastic neural networks with a large variety of dynamics, architectures, and plasticity rules: these artificial systems are composed of inputs, outputs, and plastic components that change in response to experiences in an environment. These systems may autonomously discover novel adaptive algorithms, and lead to hypotheses on the emergence of biological adaptation. EPANNs have seen considerable progress over the last two decades. Current scientific and technological advances in artificial neural networks are now setting the conditions for radically new approaches and results. In particular, the limitations of hand-designed networks could be overcome by more flexible and innovative solutions. This paper brings together a variety of inspiring ideas that define the field of EPANNs. The main methods and results are reviewed. Finally, new opportunities and developments are presented.

What problem does this paper attempt to address?

The core problem that this paper attempts to solve is to design and evolve neural network systems capable of autonomous learning. Specifically, Evolved Plastic Artificial Neural Networks (EPANNs) aim to generate neural networks with adaptive capabilities by simulating the natural evolution process. These networks not only possess innate properties but can also change and learn according to different experiences in the environment. The research objectives of EPANNs include: 1. **Autonomously create learning systems**: One of the goals of EPANNs is to be able to design and create learning systems autonomously without direct human intervention. 2. **Start learning from scratch**: Research on how to make neural networks start learning from a state without any prior knowledge and gradually develop complex behaviors. 3. **Restore performance under unknown conditions**: Explore how to make neural networks maintain or restore their performance when encountering unseen environmental conditions. 4. **Test the computational advantages of specific neural components**: Verify the impact of certain neural network components (such as different neuron types, dynamic characteristics, network architectures, and plasticity rules) on computational power through experiments. 5. **Propose hypotheses for the emergence of biological learning**: Derive hypotheses about how biological learning mechanisms evolve through the study of EPANNs. To achieve these goals, EPANNs combine the concepts of evolutionary algorithms and lifelong learning, simulating the processes of evolution, development, and learning in nature. This method not only helps to understand the formation mechanism of biological intelligence but may also promote the development of artificial intelligence technology, especially in terms of openness, autonomy, and creativity.

Born to Learn: the Inspiration, Progress, and Future of Evolved Plastic Artificial Neural Networks

Evolving plastic neural networks with novelty search

Evolving Self-Assembling Neural Networks: From Spontaneous Activity to Experience-Dependent Learning

Do What Nature Did to Us: Evolving Plastic Recurrent Neural Networks for Task Generalization

Learning the Plasticity: Plasticity-Driven Learning Framework in Spiking Neural Networks

A New Evolutionary System for Evolving Artificial Neural Networks

Evolving interpretable plasticity for spiking networks

Protein pathways as a catalyst to directed evolution of the topology of artificial neural networks

Growing Artificial Neural Networks for Control: the Role of Neuronal Diversity

Evolving Novel Gene Regulatory Networks for Structural Engineering Designs

Neuroevolution in Deep Neural Networks: Current Trends and Future Challenges

Explaining the Neuroevolution of Fighting Creatures Through Virtual fMRI

Towards Self-Assembling Artificial Neural Networks through Neural Developmental Programs

Ensemble Structure of Evolutionary Artificial Neural Networks

Recent Advances at the Interface of Neuroscience and Artificial Neural Networks

Towards the Neuroevolution of Low-level artificial general intelligence

Brain-inspired learning in artificial neural networks: a review

Evolutionary Artificial Neural Networks That Learn and Generalise Well

Neuromorphic Processing and Sensing: Evolutionary Progression of AI to Spiking

Evolutionary neural networks for deep learning: a review

Evolving the Behavior of Machines: From Micro to Macroevolution