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

**Authors:** Yurio Windiatmoko, Ahmad Fathan Hidayatullah, Ridho Rahmadi  
**arXiv:** [2009.12341](https://arxiv.org/abs/2009.12341)  
**Date:** 2020-09-25  
**Summary page:** https://www.summarizepaper.com/en/paper/2009.12341v1/

## Key points

- Smart systems powered by AI have been developed to assist humans in various tasks
- Chatbots in universities have become increasingly prevalent with recent technological advancements
- Chatbots provide immediate access to accurate information, similar to customer service inquiries
- Chatbots are functional in providing assistance for curriculum information, admission procedures, scheduling details, student grade updates, and additional features such as worship schedules and weather forecasts
- Deep Learning models using RNN and LSTM are used to develop these chatbots
- The RASA framework is integrated into the LSTM-based chatbot model
- Facebook is chosen as the platform for deploying these chatbots due to its wide user base in Indonesia
- The chatbot focuses on case studies within the Magister Informatics FTI University of Islamic Indonesia campus
- This research represents an initial stage development using simulated data that adequately simulates real-world scenarios
- The authors of the study provide further insights into the implementation and benefits of AI-powered chatbots in university settings.

## Summary

Smart systems powered by Artificial Intelligence (AI) have been extensively developed to assist humans in various tasks. One such application is the use of chatbots in universities, which has become increasingly prevalent with recent technological advancements. Particularly during the ongoing pandemic, where face-to-face interactions are limited, college students and prospective candidates often require immediate access to accurate information, similar to customer service inquiries. Chatbots have proven to be functional in providing assistance for a range of purposes including curriculum information, admission procedures for new students, scheduling details for lecture courses, student grade updates, and additional features such as Muslim worship schedules and weather forecasts. These chatbots are developed using Deep Learning models that replicate human intelligence through specific training schemes. The underlying technology behind this type of Deep Learning is Recurrent Neural Networks (RNN), specifically utilizing Long Short-Term Memory (LSTM) which efficiently manages memory allocation by removing unnecessary data. The RASA framework is integrated into the LSTM-based chatbot model. The choice of Facebook (FB) as the platform for deploying these chatbots stems from its wide user base in Indonesia, where FB users constitute approximately 60.8% of the country's population. It should be noted that this particular chatbot focuses solely on case studies within the Magister Informatics FTI University of Islamic Indonesia campus. This research represents an initial stage development using simulated data that adequately simulates real-world scenarios. The authors of this study include Yurio Windiatmoko, Ahmad Fathan Hidayatullah, and Ridho Rahmadi who published a paper titled "Developing FB Chatbot Based on Deep Learning Using RASA Framework for University Enquiries" which provides further insights into the implementation and benefits of AI-powered chatbots in university settings. Overall, this detailed summary highlights how AI-driven chatbots have become indispensable tools in universities by providing timely and accurate information to students and candidates while adapting to the challenges posed by the pandemic and the need for remote interactions.