Investigating Speech Features for Continuous Turn-Taking Prediction Using LSTMs | Read Paper on Bytez