Using Bot for searching or IVRS or talking with Chatbot is part of our daily routine. It is really amazing to understand the depth of technology behind it.
‘Bot’ and ‘Chatbot’ are two terms regularly used in digital world and have already become an indispensable part of our lives.
Bot is derived from ‘Robot’ and is a software that performs an automated task over the internet. The task is one that is repetitive, mundane and boring, which no human can endure for long. A typical example of its application is our day-to-day use of search engine. At our command of the search engine, bot systematically browses the web and throws the result pages. This is a ‘good bot.’ There is ‘malicious bot’ designed to steal information like passwords or infect viruses.
Mention must be made of ‘Captcha.’ Most of us are familiar with this term which invariably butts in after the password when we want to open some websites. Captcha is an acronym for ‘Completely Automated Public Turing test to tell Computers and Humans Apart.’ This represents a computer system intended to distinguish human from machine input. This is done by asking the user to type a combination of letters and numerals from an image, mostly a distorted one. This type of Captcha provides at least three roadblocks for a potential malicious Bot. First the characters from an image have to be physically put in. Second is to ‘recognise’ the wavy or distorted characters. Third is the ‘segmentation’ test to distinguish the characters when they are crowded together with no space in between. All these pose a challenge to a ‘malicious’ Bot and much more so when all three are present.
Most of us have the experience of contacting call centres of various organisations over our mobiles. We are met with pre-recorded greetings and then guided to choose an option from a set menu, all without a live operator. This is made possible through Interactive Voice Response System (IVRS) technology that allows a computer to interact with clients through use of input via the keypad of the mobile. For example, to book household gas refill, IVRS verifies consumer number and distributor name, confirms need for refill, books and reads out booking number; all without an operator and on 24×7 basis. IVRS is generally used for handling large volumes of calls with only simple interactions.
However, IVRS has its disadvantages. Many of us have felt exasperated when, for a simple query, we are routed through endless options until directed to press ‘0’ or ‘9’ to talk to someone who will understand our query in a precise manner. Chatbot offers a vastly improved experience for such a situation.
Chatbot is a modern upgraded version of IVRS. It is a computer programme that enables communication with users through the internet in a way that simulates real-life conversation. It allows users to interact in plain English and provide suitable responses such as weather reports, traffic situation, sports scores…
Amazon’s Alexa and Apple’s Siri are typical examples. When questioned, the Chatbot responds based on the knowledge database available to it in the computer to which it is linked. If the conversation goes on unfamiliar lines, it responds ‘I don’t understand your question.’
Working of Chatbot
It essentially works on two technologies – Speech Recognition (SR) and Natural Language Processing (NLP). Artificial Intelligence (AI) and Machine Learning (ML) are integral part of the total system for it to keep improving its response with greater usage.
SR does the task of converting human speech into its corresponding textual form. SR technology has to be effective in handling different scenarios. Hence Chatbot is fed with a big database of voice samples of large group of people. A powerful hardware with speech recognition algorithm ensures that the query or command is faithfully conveyed to servers for processing.
Understanding the intent of the query or command is important. There are number of ways of phrasing the query or command for a particular situation. Hence intent analysis becomes important. The NLP algorithm is aided by a big dataset capturing variations of the same sentence. The relevant ‘entities’ are extracted, which are the ‘key terminologies’ utilised in the query is done by the NLP algorithm. With this combined output, the server accesses the relevant website which is played back to the user.
Using IVRS or talking to ‘Siri’ is part of our daily routine. It is really amazing to understand the depth of technology behind it. We have had just a glimpse.