- Retrieval-based model – For a retrieval-based model, we use a corpus of predefined responses and match them for new contexts based on similarity with our responses
- Generative model – A generative model chatbot doesn't use any predefined corpus and can generate completely new sentences.
Generative models are able to generate more proper responses that could have never appeared in the corpus, while retrieval-based models enjoy the advantage of informative and fluent responses because they select a proper response for the current conversation from a repository with response selection algorithms. The aditional advantage of using a retrieval-based model is the fact that all unsuitable responses because of grammatical or ethical issues can be easily removed. However, your model will be limited only to the responses defined in the corpus. Generative models, oppositely, can adapt better to surprising demands and questions from customers.