A wide range of natural and engineering systems exhibit extreme events; i.e., spontaneous intermittent behavior manifested through sporadic bursts in the time series of their observables. Examples include ocean rogue waves, intermittency in turbulence, extreme weather patterns and epileptic seizure. Because of their undesirable impact on the system or the surrounding environment, the real-time prediction and mitigation of extreme events is of great interest. In this talk, I discuss some recent advances in the quantification and prediction of extreme events. In particular, I introduce a variational method that disentangles the mechanisms underpinning the formation of extreme events. This in turn enables the data-driven, real-time prediction of the extreme events. I demonstrate the application of this method with several examples including the prediction of ocean rogue waves and the intermittent energy dissipation bursts in turbulent fluid flows. I will also discuss future research directions in data-driven prediction and control of extreme events.