Mass public quarantining, colloquially known as a lock-down, is a non-pharmaceutical intervention to check spread of disease and pandemics such as the ongoing COVID-19 pandemic. We present ESOP, a novel application of active machine learning techniques using Bayesian optimization, that interacts with an epidemiological model to arrive at lock-down schedules that optimally balance public health benefits and socio-economic downsides of reduced economic activity during lock-down periods. The utility of ESOP is demonstrated using case studies with VIPER, a stochastic agent-based simulator that we also propose. However, ESOP can flexibly interact with arbitrary epidemiological simulators and produce schedules that involve multiple phases of lock-downs.