Opportunities in Neural Networks for Industry 4.0

Today, we are experiencing the Fourth Industrial Revolution, known as “Industry 4.0.” Like previous industrial revolutions, it has impacted businesses, governments, and human society. Supported by technologies such as the Internet of Things (IoT), cyber‐physical systems, and cognitive computing, Industry 4.0 has improved industrial processes by increasing economic gains and the safety of employees. In the cognitive computing context, neural networks arise as prominent tools. Their capability of modeling complex phenomena made them the state of the art of many tasks, for example, computer vision, natural language processing, and time‐series modeling. Those tasks are relevant to many industrial processes, making opportune the use of neural networks in several applications. A meaningful example of an opportunity is using neural networks for intelligent fault detection. Once the network learns the patterns related to the normal operation of industrial machines, it can recognize defects in product characteristics and machinery operation, among others. It can reduce costs related to defective products and improve the planning of more efficient maintenance routines. Another example regards safety issues. One can use computer vision systems based on neural networks to detect people in unsafe areas, such as the regions near industrial robots, to avoid potential accidents. In this chapter, we discuss those and many other opportunities to improve processes in Industry 4.0 with neural networks. Furthermore, we present two successful cases of improvement in industrial processes using neural networks: (i) detecting defects in sanitary ware with deep learning and (ii) detection of anomalies in embedded systems using electrical signatures.