Definitions, history, mathematics and EU regulation of artificial intelligence, with research examples from finance
|
Artificial Intelligence Intro (pdf)
|
9.81 M |
According to a conjecture of 1956, any aspect of learning can in principle be described so precisely that a machine can simulate it. Current artificial intelligence has three drivers, data, mathematics and computing power, and its mathematical core is the neural network as a function, with the universal approximation theorem as the main result. The history of the field has two AI winters before the deep learning period. Research examples from finance are credit risk models, peer-to-peer lending, reinforcement learning, generative adversarial networks and blockchain. Explainable AI is needed because the combination of variables into a prediction is opaque. In the European approach to trustworthy AI, systems are sorted into four risk levels and are measured against seven requirements.