Domain-Independent, Task-Oriented Chatbot Creation and Conversation Policy Management Framework
-
Tolga Çekiç [1] ; Yusufcan Manav [1] ; Enes Burak Dündar [1] ; Osman Fatih Kılıç [1] ; Onur Deniz [1] ; Seçil Arslan [1]
-
[1] YapıKredi Teknoloji, Turkey
-
- Localización: Proceedings of the Workshop on Hybrid Intelligence for Natural Language Processing Tasks (HI4NLP 2020) co-located with 24th European Conference on Artificial Intelligence (ECAI 2020): Santiago de Compostela, Spain, August 29, 2020 / coord. por Pablo Gamallo Otero, Marcos García González, Patricia Martin-Rodilla, Martín Pereira Fariña, 2020, págs. 35-39
- Idioma: inglés
- Enlaces
-
Resumen
In this paper, we present a chatbot creation framework that can help people with no technical expertise to design and cre- ate chatbots for any domain. This framework enables the creation of highly customizable chatbots that can range from simple question answer systems to chatbots that can handle more complex dialogue flows. In order to be domain independent we created a general Turk- ish language model using ELMo architecture and intent detection models for chatbots are trained using embeddings generated via this language model. Additionally, in order to make conversations more seamless and cohesive, dialogue act classification is integrated into conversation policy management. The framework also includes an additional tool that allows monitoring of past chatbot conversations and provides analytic tools supported by clustering algorithms.
