Machine Learning and Neural Networks

Neural networks are a mathematical model that directly replicates the way the human brain works. A neuron has several inputs and one output. The combination of neurons to form a network is a fantastic construction. Neurons are arranged in layers. Each layer consists of several (many) nodes that are connected to each node of the next layer. The nodes and connections have certain values that are multiplied together and added at the receiving node. If the sum exceeds a given threshold value, an output signal is emitted; if not, no value is emitted. Neural networks are a mathematical model that directly replicates the way the human brain works. A neuron has several inputs and one output. The combination of neurons to form a network is a fantastic construction. Neurons are arranged in layers. Each layer consists of several (many) nodes that are connected to each node of the next layer. The nodes and connections have certain values that are multiplied together and added at the receiving node. If the sum exceeds a given threshold value, an output signal is emitted; if not, no value is emitted. You have to imagine such a system in practice. A technical assistant takes the recordings, and a few minutes later, the attending doctor has the diagnosis on the table and has time to deal with the patient. In addition, the diagnoses are coupled with information on medication and therapies. And this applies to all imaging diseases (ultrasound images, X-rays, ECGs, and EEGs). Images of the retina are also interesting. This is the only place where the nervous system is visible from the outside. Dementia, Alzheimer’s, and Parkinson’s can be recognized immediately. Any doctor can be connected to this system, not just specialists. The system can also be easily replicated, which can benefit countries with a less developed healthcare system. Everyone can work at the highest possible level. Neural networks have only been used in the last 15 years; before that, computers did not have the necessary power. It is very surprising why these systems are not being introduced more quickly—the computers, the programs, and the Internet are already available! An extremely useful system comes from the Cologne-based company DeepL. They trained neural networks to translate texts written in different languages. At the moment, 31 languages are offered, resulting in 961 different combinations. What used to require highly qualified training can now be done by anyone. Artificial intelligence is often referred to as a job killer because of such phenomena, but this has happened many times in history; the railway has also replaced the stagecoach, etc. You just have to be prepared for such developments in good time. Learning a foreign language supports intellectual and cultural education, but is no longer necessary for business relationships.