Demystifying AI: Understanding the Basics and Breaking Down the Jargon

Artificial intelligence, more commonly known by its acronym, AI, has been rapidly incorporated into our everyday lives whether we realise it or not. In just a decade, the advances and improvements done to this technology have been to such a degree that numerous businesses are now adopting AI into their business operations. 

The field of AI combines elements from different and varied disciplines including but not limited to: computer science, cognitive science, psychology, and game theory. Because of this reason, understanding AI comes in hand with learning technical terms that may be confusing beforehand. But don’t worry, that’s what this article is for!

Before we get into high-complexity AI-related concepts, we’ll cover basic terms so there are no loose ends further in the process. 

Artificial intelligence

AI is the field that uses computer systems and datasets to simulate, as the name suggests, human intelligence and be able to carry out problem-solving tasks. The AI field is of quite an extensive width, incorporating sub-fields of machine learning and deep learning (terms we’ll go through as well later in the article), which consist of algorithms whose purpose is to create specialised systems that can make predictions and classify data based on the inputs they receive. 

Algorithm 

An algorithm is a set of rules or processes that the computer must follow to make calculations, automated reasoning, and process data. In other words, an AI algorithm serves as a code that instructs the computer how to learn to function independently. You could say that without algorithms, there’s no AI. So, that’s how you can tell how big of a role these play on the field.

Natural language processing

Also referred to as NLP, it is the branch of artificial intelligence which gives machines the ability to understand written text and spoken words resembling the way human beings communicate. Machines translate human language into numeric information by combining computational linguistics, statistical, machine learning, and deep learning models. In turn, they’re able to carry out communication tasks in real time, such as translate text to different languages, respond to voice commands, summarise in a quick manner lots of text, and many more!

You actually might have encountered NLP when making use of virtual assistants like Siri and Alexa. These use NLP to respond to people’s questions and instructions. Another common example where NLP is applied is in chatbots.

Chatbot

Chatbots are computer-based programs designed to drive conversations with human users. These are commonly used by online businesses to assist them with a variety of duties, ranging between customer service, sales, marketing, email, process management, and many others. 

For example, customer service chatbots can help answer frequently asked questions, look after enquiries, streamline messages…you name it. Chatbots enhance efficiency and open the possibility of providing support 24/7 so your team can focus your efforts in areas that require more attention.

Machine learning

Machine learning is another branch of artificial intelligence that focuses on the way that the system uses and processes data. It is able to learn and recognise patterns on its own thanks to its computer science and statistics models. In other words, the machine is capable of performing specific tasks without the need of human interference. In this way, it recognises patterns in the data which helps make future predictions as soon as new data is processed. 

In general, the learning process of these algorithms can either be supervised or unsupervised, depending on the data being used to feed the algorithms. We’ll cover the differences between these two in another article so you can dive into it a little bit deeper. But the main thing you have to understand is that there are two different ways through which the machine can learn new data.

Machine learning's greatest benefit is its ability to swiftly apply a massive amount of data to excel at many different tasks, including facial recognition, speech recognition, object recognition, translation, and many more. 

You can find machine learning being implemented on self-driving cars, image recognition applications, automation of manufacturing processes, and supply chain management.

Deep learning

Deep learning is a specialised subset of machine learning whose algorithms are mathematically more sophisticated (than those of machine learning). More specifically, the algorithms used in deep learning analyse data through a logical structure that mimics humans’ learning process. This can be obtained through either supervised or unsupervised learning. 

The algorithms used in deep learning are structured in different layers that make up an artificial neural network (ANN). The ANN design was inspired by the human brain’s biological neural network, which resulted in a learning process much more complex than that of standard machine learning models.

Deep learning has been applied in a wide range of industries. Some examples of its applications include: fraud detection, self-driving cars, visual recognition, among others. In the healthcare sector specifically, deep learning has been widely used to develop predictive models for disease diagnosis, prognosis, and treatment recommendations. It has also been applied in new imaging techniques, such as MRI and CT scan image reconstruction. 

Neural network

As we’ve just explained, neural networks used in deep learning resemble those used in the human nervous system and brain in order for the AI to carry out its functions. For this reason, it’s no surprise that these networks are also known as biologically inspired computational models. 

The ability of AI to solve complex problems is attributed to the neural network, as it allows breaking down data in different layers. Overall, neural networks are applicable to numerous fields such as business, data mining, medicine, telecommunications, pure sciences, and operations management. 

Data mining

In artificial intelligence and machine learning, data mining explains the process of filtering valuable information from a large database. Its main purpose is to inspect all the raw information to extract useful information from it. To achieve this, data mining uses statistical methods to identify trends and form data relationships within the database that’s being analysed. 

The resulting data from data mining can be used for applications such as risk monitoring, business management, production control, market analysis, engineering, and science exploration.

Black box

An AI black box refers to a system whose inner workings are hidden and ultimately unknown to the creator(s). The creators can feed the system with input and obtain output in exchange. However, they won’t be able to examine the system’s code or the logic behind the output.

A black box can be used to store big data. For example, some AI developers may put their model or the training data used for the model in a black box to protect their intellectual property. 

The irony is that AI, itself, can act as a black box considering that it may behave and evolve in unexpected ways that may be hard to fully understand even for the system’s programmers.

We hope that this article has provided you with clarity about some common terms in AI. Even if you were already familiarised with some of the concepts we’ve just covered, it’s always good expanding or updating your knowledge in the field because trust us, AI will continue evolving. Keep an eye out on our AI Hub, as many more articles are yet to come where we'll be covering more information on AI!