Chapter I.Introduction1.1 Introduction to ChatbotsChatbots are “human-computer dialog online systems with natural language.” The concept of chatbot firstly, has been mentioned by Alan Turing (1950), who made the question “Can machines think?” 1. Chatbots are computer programs that early were programmed and designed for academic purposes, (since Alan Turing) but lately, they have become more sophisticated and are giving opportunities for intelligent communication, causing the push of boundaries by being considered more human-like. Chatbots are becoming an important part of many people’s lives because its usage is widely spread for business purposes where people use Chatbots to facilitate their job on daily basis.
The dialogue of “written characters” in a modern and digital area is a difficult process because in these days there are various versions of online chats and many companies or businesses use this computer program to increase their profits (based on the company) and chatbot is mostly used in dialog systems for instant messaging platforms considering, and mentioning here the launch of Chatbot platforms by Facebook 2, Kik 3, Slack 4, Skype 5, WeChat 6, Line 7, and Telegram 8. In 2016 (September) 30,000 bots were hosted by Facebook Messenger which had 34,000 developers on its platform 9. In 2016 (August) The Kik Bot Shop announced that 20,000 bots which were created on its platform “had exchanged over 1.8 million messages” 10. This thesis is made for a chatbot project (EPA) I created for an Albanian company (EasyPay). This chatbot is a program that uses textual methods to give responses to clients using the Messenger platform.For this thesis, the most important thing is on the process of implementation and features of my chatbot, and of course, the general information will be very useful to answer some questions that arise for the usage of online chats.
I chose Chatbots as my diploma thesis because the world is in the early stages of this new phenomenon and also I noticed a gap of information about the creation and purpose of Chatbots.1.2 Main questions The main questions this thesis tries to answer are:- What is a chatbot?- What kind of company would use chatbot?- What are some of its features?- What are the main requirements for chatbot implementation?- What are some differences between chatbot and other online chats?- What are some advantages and disadvantages of chat implementation offer to a company or simple users?1.3 Objectives and Purposes The objectives of this thesis are to present a well-designed, programmed and well-documented project and to show its useful usage, and of course, challenges that this chatbot offers to its user. Also, to show that artificial intelligence is beginning to gain terrain as one of the most developed fields. One of the main purposes of this thesis is to give information on implementing an online chat and give the advantages of it.1.4 Artificial Intelligence (Conversations)The conversation that the user makes with chatbot is called artificial conversation, and the interaction is mostly between a software entity and a human, that exchange text-based (simple short text) conversations on chat.
One major task for a chatbot is the programming and the design of an architecture which will have a better performance and “will do the work”.1.4.1 NLP and Machine LearningThe area of Artificial Intelligence and linguistics involves Natural Language Processing (NLP) which are some techniques for the analysis and processing of written language.
These analyses include search, extraction, and retrieval of information etc.The area of Artificial Intelligence (AI) includes some techniques that are used for solving problems that can be used for various problems like detection or prediction, even NLP, which are called with the term Machine Learning (ML). ML includes an approach or extension called Deep Learning that includes Neural Network which uses algorithms that deal with NLP tasks.The link between NLP and ML is in the conversion of an unstructured text into a structured one, because there are lots of techniques that are needed for this case especially deep learning techniques like K-means, Regression etc.A very powerful component is the analyzation of the text manually (how a person would analyze it), and then it’s the ML approach that will do the process automatically.