Fig. 8: Testing the limits of the transfer learning: convolutional encoder-decoder (CED) model with cross-training of the latent space applying limited portions of acoustic emission (AE) data from a single slip cycle. | Nature Communications

Fig. 8: Testing the limits of the transfer learning: convolutional encoder-decoder (CED) model with cross-training of the latent space applying limited portions of acoustic emission (AE) data from a single slip cycle.

From: Predicting fault slip via transfer learning

Fig. 8

The model input signal is acoustic emission (AE in red, left axis), and the output signal is the friction coefficient (μ in blue, right axis). The portion of the time series shaded light green shows the post-failure data used for the transfer learning exercise and shaded light blue shows the pre-failure data.

Back to article page