What is computer vision, a large language model, and how machine learning works – we explain trending tech terms in simple words.

Almost anyone can try ChatGPT artificial intelligence, Bing chatbot, or Midjourney image generator for work, education,n or just for fun. But not everyone understands how modern technology works.

Liga. Tech explains the basic terms that will help you better understand artificial intelligence.

Machine learning

The author of this term is called the researcher Arthur Samuel, who, while working at IBM, created the world’s first program capable of self-learning. At his request in 1963, checkers master Robert Neely competed against a computer until he lost to a machine.

Machine learning is behind the work of chatbots, translation programs, and even the shows that Netflix offers you. In recent years, machine learning has become an essential part of artificial intelligence, which is why it is often used as a synonym for AI.

Neural networks

Neural networks are the foundational technology for AI. Experts advise thinking of them as the equivalent of the steam engine in the first industrial revolution – a general-purpose technology that spans many different industries and has many different use cases.

The advent of neural networks began with an attempt to model the brain of animals, consisting of millions of simple neurons, each of which is connected to several others. Each neuron is very simple, but together they can learn to perform complex tasks. The same goes for artificial neural networks, which rely on training data to improve their accuracy over time.

When these learning algorithms are fine-tuned, they become powerful tools to classify and cluster data at high speed. Speech or image recognition tasks can take minutes or hours – it would take a human incomparably longer.

Big language model

Large language models are artificial intelligence systems that understand and generate text. The most famous example is ChatGPT from OpenAI.

These language models are typically tens of gigabytes in size and learn from huge amounts of textual data, sometimes on a petabyte scale. For example, Google’s PaLM language model uses “high-quality web documents, books, Wikipedia, conversations, and GitHub code” to develop language understanding. The larger the language model, the better the result. So experts predict that they are not yet close to the limit of their capacities.

Deep Learning

Deep learning is a field of artificial intelligence that allows machines to master the tasks people usually perform. This machine-learning technique allows computers to learn from human examples, which helps automate various processes.

For example, it is a key technology for self-driving cars that allows them to recognize a stop sign or distinguish between a pedestrian and a light pole. Deep learning technology is at the heart of everyday products and services such as digital assistants, voice-enabled TV remote controls, and more.

Data Science

Today, Data Science is used in almost every industry. Businesses can use it to make informed product development and marketing decisions. Governments are also using Data Science to improve the efficiency of public service delivery.

Why does AI need Data Science? Before a computer program tries to learn from the data, it is useful for a human (or a data analysis program) to learn from it. This allows them to clear the excess and highlight the importance. This intervention helps to improve the AI ​​learning process.

computer vision

Computer vision requires a lot of data and is used in industries ranging from energy to automotive. For example, to teach a computer to recognize car tires, it needs to provide a huge number of images of tires and objects related to them. Then he will be able to understand the differences between them and determine, in particular, tires without defects.

Intelligent Robotics

Intelligent Robotics is an industry that uses artificial intelligence to improve collaboration between people and devices. AI helps robots adapt to dynamic situations and communicate naturally with humans.

The choice that such a robot must make is related to the intelligence built into it using a machine or deep learning, as well as to the input data received by the robot during operation.

Guided Machine Learning

Supervised learning is a form of machine learning that does not function on its own, but requires human intervention. Data is fed into the machine and the process is controlled by a human while the computer works on a particular result.

Machine learning without supervision

Unsupervised Learning – This type of machine learning does not require human intervention or allows the machine to conclude on its own based on patterns found. Unsupervised machine learning aims to discover previously unknown relationships in data. For example, it can be used to determine the target market for a completely new product that your business has never sold before.