Artificial intelligence

Understanding Artificial Intelligence and the breath and depth of its capabilities, can be grasped by diving into today's and upcoming use cases, or alternatively starting bottop-up from its algorithms.

AI is used in Natural Language Processing (NLP) e.g. to come to the level of understanding of human written and/or spoken language by a compter e.g. a chatbot. AI is at the basis of facial recognition or within a wider domain of machine vision. A selfdriving car is equipped with such capabilities and many other AI-driven applications; just think of the many decisions to be taken in realtime or quasi-realtime (e.g. predictive maintenance).

From automotive applications, we may jump to industrial applications with anomaly detections, but equally to healthcare. Remote monitoring on health indicators will allow to predict chances on disorders of some sort or advise on taking medicine. Also in doing business there are many indicators where AI is pivotal in quick decisions, targetted actions, personalisation, etc.

From algoritmic point of view,  deep learning using neural networks are popular today. They are very effective, but may lack explainability. Similar for reinforcement learning and several supervised learning algorithms (e.g. recommenders). Will causal networks allow for a next move here?