Mesterséges neurális hálózatok az állatitermék-előállításban

The rise of artificial intelligence (AI) is not going unnoticed in the agricultural sector. The processing of the large amounts of data (’big data’) generated in animal production is increasingly being done using artificial intelligence, particularly machine learning (ML). Machine learning is a branch of AI, in which algorithms are automatically trained to solve a task of interest using a given dataset. There are several sub-areas of ML, of which we focus on artificial neural networks (ANNs), the most successfully used in agriculture. The basic units of an ANN are artificial neurons. These are connected to each other similarly to synapses in the brain, forming a network. ANNs can be considered complex mathematical models that can make predictions from given data after a learning process, taking into account millions of parameters. Because they are pretty flexible, these networks have a wide range of applications in many fields. One such field is a subset of agriculture, namely animal production. In our work, we outline the general structure and operation of ANNs. We provide insight into the metrics widely used to indicate the accuracy of prediction and their calculation methods. Possible applications are illustrated with examples specifically from the field of food production. The wide range of applications is illustrated by the fact that the works cited also respond to the challenges faced by aquacultures and beekeepers, in addition to the problems of cattle, pig and poultry farms. Despite their many good features, ANNs cannot solve all problems, regardless of type. Therefore, in our work we also concern about the limitations of the method. Our work contributes to the definition of artificial intelligence, machine learning, and artificial neural networks in the context of agriculture.