Application of artificial neural network in nuclear spin echo experiments

The possibility of artificial neural network application to detect nuclear spin echo signals under conditions when the echo amplitude is comparable to the amplitude of the noise is demonstrated. Data obtained by superimposing the model echo signals of a Gaussian form on experimentally recorded noise signals is proposed to use for training the neural network.