Machine learning is related to many other fields in its defining principles as well as delivery techniques. Here, we will examine its most important relationships with other fields of study. ## Artificial IntelligenceMachine learning emerged as a separate field of study during scientists’ pursuit of developing artificial intelligence (AI) in machines. They aimed to make machines learn from available data, using methods like neural networks, linear statistical models and probabilistic reasoning. However, by 1980s, expert systems were seen to be the right approach to achieve AI and all other approaches were dropped. By 1990s, machine learning was restructured as a separate field and focused on solving practical problems rather than acquiring AI. ## Computational StatisticsNew principles of machine learning borrowed heavily from computational statistics. Two models – data model and algorithmic model – used in machine learning have their roots in computational statistics. Michael I Jordan, an American scientist and researcher in both machine learning and AI, has gone as far as to suggest that the overall field of machine learning and statistics be called data science. ## ProbabilityMachine learning utilizes probability theory for studying and exploring algorithms for pattern recognition, the building block of machine learning. These algorithms are then used to build models from available data to predict machine behavior in case of new inputs. Machine learning used loads of digitized information readily available through the Internet. ## Data MiningMachine learning and data mining both utilize similar methods to achieve different goals. Machine learning focuses on making predictions based on available data whereas data mining focuses on discovering unknown aspects of available data. Machine learning uses many techniques employed by data mining. ## Mathematical OptimizationIn simple terms, optimization can be defined as finding the best solution out of all solutions for any problem. In machine learning, predictions are made on the basis of available data. Mathematical optimization techniques can be used to deliver on the right theory to be used. Thus optimization and machine learning are closely related to each other. |