Artificial Intelligence and Chatbots: Teaching the machine to think

Understand what is Artificial Intelligence for chatbot, why is not the same thing as Machine Learning, what are the available chatbot technologies and trends in this new Market

Even if you do not realize it, more and more Artificial Intelligence is present in your routine. Smartphones, cars with parking sensors, smart houses, banks, restaurants, call centers. And it doesn’t stop there: increasingly the AI gains complexity to bring many other technological applications.

But where does this complexity come from? This area of computing uses computational methods and devices to make the machine “learn” to use the rational capacity of the human being. It’s not futuristic film stuff: machines can actually learn through experiences.

Surfing the wave of machine learning, marriage between Artificial Intelligence and chatbots will be an increasingly strong trend to provide better experiences for users and bring value solutions to companies. Following the series about bots, MJV brings another article to better understand the subject of the moment.

But what is Artificial Intelligence?

Artificial Intelligence is nothing more than the ability of a computer to perform functions as a human would. This process includes some steps, such as reasoning skills, learning and pattern recognition.

The concept of Artificial Intelligence came a long time ago. Some believe that it may have been idealized at the time of the Greek philosophers. However, its breakthrough came in the nineteenth century – with English mathematician Alan Turing and his machines.

Machine Learning x Artificial Intelligence

There is a confusion between the concept of Artificial Intelligence and Machine Learning, but these are not the same thing. Machine learning is a feature of Artificial Intelligence and is characterized as a data analysis method and standards that automates the construction of improved models.

It’s Machine Learning that will allow the machine to make decisions and become able to make predictions based on algorithms. It is classified as supervised learning, unsupervised and reinforcement. It can be said that this is a new programming paradigm, because it is a system trained rather than programmed (such as traditional computing).

AI concepts

Intention

It’s what the user means by the message he sent to the bot, the main idea of a sentence. It is through intentions that we can understand the actions that the client wants to perform. In the context of chatbots, intentions are characterized by verbs.

Example: in the sentence “I want to download an ebook”, the intention would be the word “download”.

Entity

It is about what the intention refers to, that is, they are the important parameters for the actions that the user wants to perform. Entities complete the meaning of intentions. In the context of chatbots, entities are usually characterized by nouns, adjectives, services, and products that are part of the customer’s business.

Example: in the sentence “I want to download an ebook”, the entity would be the word “ebook”.

NLP ou PLN

The abbreviation for Neuro-Linguistic Programming (NLP) means Processamento de Linguagem Natural (PLN) in Portuguese. This is a subarea of artificial intelligence that studies the ability and limitations of a machine to understand the language of humans. The goal of the PLN is to give the machine enough subsidies so that it has the ability to understand and compose texts. That is, it is what enables the interaction between man and machine.

What technologies are available?

Watson

Robust IBM’s artificial intelligence is a cognitive oriented service platform for business. IBM’s goal in creating Watson was to help professionals and companies build their own cognitive systems and thereby improve processes and services. One of the possibilities of using Watson’s AI is customer service through chatbots.

LUIS

Microsoft’s LUIS is a machine learning service that builds natural language understanding not only in chatbots, but also in IoT applications and devices. With LUIS you can use voice recognition and create an even more sophisticated chatbot.

DialogFlow

Google’s Artificial Intelligence is actually more than a tool. It is a complete development platform designed to create and deploy conversation interfaces on IoT sites, applications and devices. With 20 compatible languages, it allows natural interactions between users and businesses through the creation of chatbots.

What are the trends for the future?

In our series of chatbots we have already shown how this tool can deliver value to the corporations taking into account the change of paradigm relationship with the customer. However, chatbots do not always have an Artificial Intelligence behind them. In some cases, bots use only a good conversational flow.

The way humans interact with each other depends on several factors. When we think of language – whether written or spoken – we need to consider culture, regionalisms, contexts, slang, and many other parameters practically unpredictable. That’s where the magic of Artificial Intelligence happens.

A chatbot that uses AI behind will be able to learn not only from the inputs of the developers and with their design conversation flow, but mainly with the users  interactions. The more he talks to his clients, the more he will come close to human conversation. A chatbot can even create compelling texts with such quality that they can be used to generate fake news and deceive Internet users – such as the GPT-2 system.

In the United States, scientists at the OpenAI NGO have stated that it will be necessary to launch this language prediction system in a reduced version. Their concern is because GPT-2 is too good at what it does. Its AI has absorbed data from eight million web pages to perform its task with mastery: predict the next word, considering all the previous words in some text.

Man or machine?

Lately we have been able to see Artificial Intelligence applied on a large scale in Science, in Medicine, in corporations and even in our day to day. And that’s just the beginning. The tendency is that, increasingly, machines will be able to think and even act like humans.

If today we have difficulty distinguishing a robot from a human during a conversation, imagine when they are more trained and improved. In addition to being useful, Artificial Intelligence – just like your chatbots – becomes more and more impressive.

To learn more about how a chatbot can help corporations, read the blog post Top 5 benefits of a chatbot for your company. And if you have already used or created a chatbot, leave your comment on the subject!