Open Access Open Access  Restricted Access Subscription or Fee Access

Intello-An Intelligent Chatbot For Replacing FAQs

Anamika Gupta, Gunjan Gupta, Arjun Malhotra, Khushboo Chitre, Ojasvi Aggarwal, Nitesh Kumar Gupta

Abstract


A user often needs to go through each question of the FAQ section of a website simply to find the answer to his solitary question. It would be much simpler if there existed a single interface through which he could simply type in his question and the answer is directly provided. There has been a long list of chatbots developed for the e-commerce and online customer service industry.

In this paper, we describe Intello, a chatbot developed by us, to answer all queries related to an academic institute. Primary aim of the chatbot is to replace FAQ section of the web site of an academic institute. Intello is suitable for an environment where user is not well conversed with the English language and hence chances of grammatical and spelling mistakes are high. Intello has been designed to store customised patterns depending on the developer unlike existing developer tools where contexts are generated automatically based on the training provided by the developer, and may not cover majority of the possibilities. Customized patterns can handle the spelling and grammar errors in an efficient manner.

Existing tools have a limitation on the number of queries answered per month because of business and technical reasons. Intello doesn’t impose any such limitation.

Intello learns from its failure. A knowledge base is created from the questions which doesn’t match a pattern. Later, learned knowledge is uploaded on a periodic basis and in a controlled manner. A user-friendly admin interface helps in uploading the knowledge base.

Intello is designed from scratch using Python and AIML. Python was used to implement the heart (the kernel). AIML was used to implement the brain (the knowledge base).

The developed chatbot was tested rigorously. Experiments were performed on admission procedure of Shaheed Sukhdev College of Business Studies, University of Delhi, India. Results reveal that Intello was able to answer accurately almost every question about the admission procedure. A machine learning file was prepared from the failure of Intello and knowledge base was updated. In future, Intello can be extended to include knowledge base related to other procedures of an educational institute like course structure, academic result, student activities, notices and announcements etc.  


Full Text:

PDF

References


Chatbot - https://en.wikipedia.org/wiki/Chatbot

Kuligowska, Karolina, Commercial Chatbot: Performance Evaluation, Usability Metrics and Quality Standards of Embodied Conversational Agents. Professionals Center for Business Research, Pg 1-16, 2015

Joseph Weizenbaum, ELIZA—a computer program for the study of natural language communication between man and machine. Communication of ACM 9, Pg 36-45, 1966

Parry, https://www.chatbots.org/chatbot/parry

A.L.I.C.E AI Foundation, http://www.Alicebot.org/, 2002

Olga Davydova, 25 Chatbot Pltforms: A Comparative Table, https://chatbotsjournal.com/25-chatbot-platforms-a-comparative-table-aeefc932eaff, 2016

Bayu Setiaji, Ferry Wahyu Wibowo, Chatbot Using A Knowledge in Database, 7th International Conference on Intelligent Systems, Modelling and Simulation, IEEE, 2016

Shah, Huma & Henry, Odette. The Confederate Effect in Human-Machine Textual Interaction, WSEAS, ISCA, May 2005

Alice Kerly, Phil Hall, Susan Bull, Bringing chatbots into education: Towards natural language negotiation of open learner models, In Knowledge-Based Systems, Volume 20, Issue 2, 2007, Pg 177-185

Building a Chatbot: Analysis & limitations of modern platforms https://tryolabs.com/blog/2017/01/25/building-a-chatbot-analysis-limitations-of-modern-platforms/

AIML Tutorial – https://www.tutorialspoint.com/aiml/

Features of Wit.ai vs Api.ai - https://www.quora.com/Features-of-wit-ai-vs-API-ai

Wit.ai https://www.wit.ai/

Motion.ai https://www.motion.ai/

Api.ai https://www.api.ai/

Abu Shawer B. and Atwell A. Different measurement metrics to evaluate a Chabot system Proceedings of the NAACL'07 Workshop: Bridging the Gap: Academic and Industrial Research in Dialog Technologies, pp.89-96, ACL, 2007

Güzeldere, G. and Franchi, S. Dilaogue with colourful personalities of early ai”. In Constructions of the Mind, SEHR, Volume 4, Issue2, 1995

Aitools.org. (2004), http://www.aitools.org/livebots/

Julia Hirschberg, Christopher D. Manning, Advances in Natural Language Processing, Science, Vol 349, Issue 6245, pp 261-266, Jul 2015


Refbacks

  • There are currently no refbacks.