Do not let the beginning trap you! On inhibition, associative creative chains, and hopfield neural networks

Creative thinking stems from the cognitive process that fosters new ideas and problem-solving solutions. Creative cognition in emerging artificial intelligence systems and neural models may reduce complexity in understanding creative cognition. Hopfield Neural Networks (HNN) is a simple neural model known for its biological plausibility to store and retrieve neuron patterns. The primary objective is to demonstrate that ideas, symbolized as patterns of ones and zeros representing clusters of neurons that synchronize their firing, can be stored within HNN and establish connections through correlation. The network can converge towards these ideas by manually adjusting specific state parameters, effectively controlling the overall network activity. When the second closest stored pattern deviated significantly from the input pattern, the networks ability to converge decreased, enabling it to connect the input and patterns. Thus, we suggest employing HNN for the first time to create a model that emulates creative thinking processes, including making meaningful links between seemingly unrelated ideas. We implemented certain modifications to the original HNN, including introducing pattern weight control, which provides a robust representation for content addressable memory and illustrates conceptual links in stored data, a step towards the larger framework of creativity. We have made progress in identifying two mechanisms that could assist in managing the dynamics of the network and the formation of associative links. These mechanisms are related to the activation threshold of the neurons and the inhibitory stimulus on the stored patterns.