Just as the cryptocurrency boom brought with it a lot of new jargon, the AI boom has brought with it a lot of new terms that we often hear but don’t necessarily understand.Professional terminology.
If you want to learn more about chatbots andLLMIf you’re wondering what the difference is between artificial intelligence (AI) and machine learning, or what the difference is between deep learning and machine learning, you’ve come to the right place. Here are 20 AI-related terms with detailed explanations.
Artificial Intelligence (AI)
In simple terms, artificial intelligence is about giving computers or machines human-like intelligence. This term is very broad and includes many different types of machine intelligence.
The current discussion focuses on tools that can create art, content, or summarize and transcribe content. Although there is still controversy over whether these tools should be called "intelligent", the term "artificial intelligence" has become widely accepted.
algorithm
An algorithm is a set of instructions that a program follows to produce a result. Common examples include search engines that display a range of results based on your query, or social media apps that display content based on your interests. Using algorithms, AI tools can create predictive models or generate content or artwork based on your input.
bias
In the field of artificial intelligence, bias refers to incorrect results caused by the algorithm making incorrect assumptions or lacking sufficient data.
For example, speech recognition tools may not understand certain English accents correctly because they have only been trained with American accents.
Conversational AI
Conversational AI is a technology that uses natural language processing (NLP) and machine learning to enable computers to understand, process, and generate human language, allowing for fluent conversations. For example, smart voice assistants such as Apple's Siri and Amazon's Alexa are typical applications of conversational AI.
Data Mining
Data mining is the process of finding patterns or trends in large amounts of data. Some AI tools use data mining to help you understand why people buy more in your store or on your website, or how to optimize your business to handle peak demand during peak hours.
Deep Learning
Deep learning attempts to mimic the way the human brain learns, using three or more neural network layers to process large amounts of data and learn by example. These layers each process their own view of the given data, which is then aggregated to draw a final conclusion.
Self-driving car software works by using deep learning to recognize stop signs, lane markings, and traffic lights: it does this by showing the AI tool many examples of a particular object, such as a stop sign, and through repeated training, it eventually is able to recognize those objects with an accuracy rate approaching 100%.
Large Language Model (LLM)
Large Language Models (LLMs) are deep learning algorithms trained on massive datasets to generate, translate, and process text.
LLMs (like OpenAI’s GPT-4) allow AI tools to understand your query and generate text input based on it. LLMs also help AI tools identify important parts of a text or video and summarize it for you.
Generative AI
Generative AI can generate art, images, text, or other results based on your input, and these results are usually powered by LLM. It has become an umbrella term for many companies that are currently using this type of AI technology in their products.
For example, a generative AI model can generate an image given some text prompts, or turn a vertical photo into a widescreen wallpaper.
Hallucinations
When AI interprets fiction as fact, we call it hallucination. Hallucination occurs when the AI is fed an inaccurate data set or has flawed training, so it outputs an answer it is certain of based on existing knowledge.
Due to the complexity of artificial intelligence, we don't always understand the specific reasons for these hallucinations.
Image Recognition
The ability to identify specific subjects in an image. Image recognition can be used by a computer program to spot and name flowers in an image, or to identify different species of birds in a photograph.
Machine Learning
Machine learning is a technology that enables computers to automatically improve their performance through algorithms and data. It does this by learning patterns and regularities from large amounts of data in order to make decisions or predictions without being explicitly programmed.
For example, an email spam filter is an application of machine learning. It analyzes the content of a large number of emails and learns how to distinguish between spam and good emails, thereby automatically filtering out spam.
Natural Language Processing
Natural language processing refers to the ability of a program to understand and process input written in human language. For example, when you ask Siri, “What’s the weather like today?” your calendar app or Siri can understand what you are saying.
Neural Networks
The human brain has many layers of neurons that continuously process information and learn from it. Artificial intelligence's neural networks mimic this structure of neurons and learn from data sets. Neural networks are systems that implement machine learning and deep learning, which ultimately allow machines to perform complex tasks such as image recognition and text generation.
Optical Character Recognition (OCR)
Optical Character Recognition (OCR) is a technology that extracts text from images. Programs that support OCR can recognize handwritten or typed text and also allow copying and pasting.
Prompt engineering
Prompt engineering is the art of designing and optimizing text prompts fed into an AI model to obtain the desired output.
In terms of AI, prompt engineering is the art of writing prompts so that the chatbot gives the most helpful responses.
Reinforcement Learning from Human Feedback (RLHF)
RLHF is the process of using human feedback to train AI. When the AI gives an incorrect result, humans show it the correct response. This allows the AI to provide accurate, useful results at a faster rate.
Speech Recognition
The ability of a program to understand human language. Speech recognition can be used in conversational AI to understand your queries and provide responses, and in speech-to-text tools to understand spoken language and convert it into text.
Token
When you input a text query into an AI tool, it breaks the text down into tokens (sequences of characters commonly found in text) which are then processed by the AI program.
For example, if you’re using a GPT model, pricing is based on the number of tokens it processes: You can calculate this number using the company’s tokenizer tool, which also shows you how words are broken down into tokens.
OpenAI says one token is roughly equal to four characters of text.
Training Data
Training data is the data that machine learning models use to learn and improve. Just like a student does exercises to master knowledge, machine learning models analyze this data to identify patterns and regularities to make predictions or classifications in new situations.
For example, if we want to train a program to recognize cats and dogs in photos, we will give it a large number of labeled photos of cats and dogs, and the model will learn how to distinguish between cats and dogs through these photos. The quality and quantity of training data directly affects the performance of the model.
Turing Test
Alan Turing was a British mathematician who is known as the "father of theoretical computer science and artificial intelligence." His Turing test (or "imitation game") was designed to determine whether a computer's intelligence is the same as that of a human.
If a human is fooled by the computer's response into thinking it was written by a human, then the computer has passed the Turing test.