There are three kinds of artificial intelligence (AI), experts systems, neural networks and genetic algorithms. These can all be helpful for scientists to solve medical problems.
A type of AI called an expert system uses rules and knowledge to replicate the decisions-making capabilities of human experts. A medical expert system, for example, might help diagnose the condition of a patient based on his or her symptoms and their medical history. Isabel Diagnosis Support System (a web-based system which assists doctors in diagnosing complex and rare diseases) is an example of such an expert system. This system is widely used by hospitals and clinics all over the globe.
The neural network, a machine learning algorithm that mimics the function and structure of the human brain, is one type of neural network. The neural network is used to analyse medical data, make predictions and/or give recommendations. Based on a person’s medical history and genetic data, a neural network can be used to predict the probability of them developing a specific disease. DeepMind Health, which analyzes medical records to improve the outcomes of patients at London’s Royal Free Hospital, is an example of such a neural network.
Genetic algorithms are a form of optimization that employs genetics and natural selection to solve a problem. Genetic algorithms can be applied to medical problems in order to determine the most effective combination of treatments for each patient, taking into account their particular characteristics and past medical history. The Oncology Expert Advisor is an example of a current genetic algorithm that is used to help oncologists select the most effective treatment options for patients with cancer based on their specific tumor characteristics.