Neuron Coverage refers to the measure of activated neurons in a neural network during testing. It assesses how many neurons respond to given inputs, providing insights into the network’s behaviour and learning capabilities.
Key aspects of neuron coverage:
- Evaluates the effectiveness of neural network training.
- Identifies areas where neurons may not be adequately utilised.
- Helps optimise neural network architecture for better performance.
High neuron coverage indicates that the neural network is effectively learning from input data, while low coverage may suggest areas for improvement in training or architecture design.