US20220121941A1 - DRIVING STOCHASTIC AGENTS TO ENGAGE IN TARGETED ACTIONS WITH TIME-SERIES MACHINE LEARNING MODELS UPDATED WITH ACTIVE LEARNING