Current State-of-the-Art (SotA) chatbots are able to produce high-quality sentences, handling different conversation topics and larger interaction times. Unfortunately, the generated responses depend greatly on the data on which they have been trained, the specific dialogue history and current turn … Continue reading →