The history of AI is often seen in terms of relative excitement and success followed by periods of disillusionment, called AI winters. The most recent, sometimes simply called 'the AI Winter' was between the mid 1990s, following the early promise and over-hyping of neural networks and early 2010s when the devlopemnt of deep neural networks led to the current high interest in AI. Some fear that over-inlfalted expections of what AI can achieve may lead to yet another AI winter, but the widespreadsuccessful practical applications of AI makes this seem less likely.
Defined on page 9