Developing Facebook Chatbot Based on Deep Learning Using RASA Framework for University Enquiries

Yurio Windiatmoko,Ridho Rahmadi,Ahmad Fathan Hidayatullah

doi:10.1088/1757-899x/1077/1/012060

Yurio Windiatmoko, Ridho Rahmadi + Show 1 more

Open Access

https://doi.org/10.1088/1757-899x/1077/1/012060

Copy DOI

Abstract

Smart systems for universities powered by artificial intelligence have been massively developed to help humans in various tasks. The chatbot concept is not something new in today’s society which is developing with recent technology. College students or candidates of college students often need actual information like asking for something to customer service, especially during this pandemic, when it is difficult to have an immediate face-to-face meeting. Chatbots are functionally helping in several things such as curriculum information, admission for new students, schedule info for any lecture courses, students grade information, and some adding features for Muslim worships schedule, also weather forecast information. This Chatbot is developed by deep learning models, which was adopted by an artificial intelligence model that replicates human intelligence with some specific training schemes. This kind of deep learning is based on RNN which has some specific memory savings scheme for the deep learning model, specifically this chatbot using LSTM which already integrates by RASA framework. LSTM is also known as Long Short Term Memory which efficiently saves some required memory but will remove some memory that is not needed. This Chatbot uses the Facebook platform because the Facebook users have already reached up to 60.8% of its entire population in Indonesia.

Highlights

Chatbot is the easiest system to access by any users, chatbot is an artificial intelligence technology for communicating through natural language with humans
This Chatbot is developed by Deep Learning models, which was adopted by an artificial intelligence model that replicates human intelligence with some specific training schemes
Deep learning is a branch of Machine Learning, inspired by the neural network of the human brain which is represented by layers on the neural network [4]. This kind of Deep Learning is based on RNN which has some specific memory savings scheme for the Deep Learning Model, this chatbot using LSTM which already integrates by RASA framework

Summary

Introduction

Chatbot is the easiest system to access by any users, chatbot is an artificial intelligence technology for communicating through natural language with humans. A chatbot is needed to answer every question quickly, correctly, and is available anytime which is expected to work for 24 hours [2] This Chatbot is helping in several things such as curriculum information, admission for new students, schedule info for any lecture courses, students grade information, and some adding features for Muslim worships schedule, weather forecast information. This Chatbot is developed by Deep Learning models, which was adopted by an artificial intelligence model that replicates human intelligence with some specific training schemes. The aim and philosophy of the initiators of the Rasa chatbot Framework are to create Machine learning-based dialogue management that provides ease of use in terms of implementation, as well as bootstrapping, by starting the stages from existing sources to create something more complex and in a more efficient way, even with minimal initial training data [5]

Literature Review

Framework

Findings

Result and Discussion

Conclusion and Future Work