Summary

In this chapter, the CNN architecture built in Chapter 9, Abstract Image Classification with Convolutional Neural Networks (CNNs), was loaded to classify physical gaps in a food processing company. The model uses image concepts, taking CNNs to another level. Neural networks can tap into their huge cognitive potential, opening the doors to the future of AI.

Then, the trained models were applied to transfer learning by identifying similar types of images. Some of those images represented concepts that led the trained CNN to identify concept gaps. Image concepts represent an avenue of innovative potential adding cognition to neural networks.

concept gaps were applied to different fields using the CNN as a training and classification tool in domain learning.

concept gaps have two main properties: negative n-gaps and positive p-gaps. To distinguish one from the other, a CRLMM provides a useful add-on. In the food processing company, installing a webcam...