The octopus approach to the Alexa competition: A deep ensemble-based socialbot

We present a novel, large-scale ensemble-based system for the Amazon Alexa Prize competition. Our system leverages state-of-the-art methods from deep learning and reinforcement learning. We carry out A/B testing experiments with real-world users and demonstrate that our approach yields substantial improvements over several baseline systems. During the competition semi-finals, our best performing system obtained a substantially higher average Alexa user score and number of back-and-forth turns compared to the average of all teams. Due to its machine learning architecture, our system is likely to improve with additional data.