Deep learning Indonesian chatbot using PyTorch for customer support automation

Abstract

The customer is the frontline as a tool for marketing products or receiving product complaints. Increasing the business process will be increasing the response rate and response time of answering the customer. A chatbot can automate the customer support system to answer various customer needs. The chatbot is software that allows human interaction with machines using natural language. The Generative Chatbot can be trained using the conversation dataset and generate a response based on input. In this research, are build the generative Indonesian chatbot model using Deep Learning. The method of Deep Learning is Sequence to Sequence with Attention using the PyTorch framework. This chatbot makes customer support automation and helps the customer support faster to answer the question. The initial result of this chatbot has a BLEU Score of 0.98.