top of page

ARTIFICIAL INTELLIGENCE

(ai)

Understanding the following key terms will help you grasp the foundational concepts of AI and its transformative potential. Whether you’re a beginner or looking to deepen your knowledge, this guide provides a solid starting point for exploring the world of artificial intelligence.

Below is a list of the key terms often used in Artificial Intelligence. 

Learning them may not make you smarter, but at least you'll know what everyone's talking about!

 

Agent - A physical or virtual entity that can act, perceive its environment (in a partial way), and communicate with others, is autonomous, and has skills to achieve its goals and tendencies

 

AI Bias - The tendency of AI systems to produce prejudiced results due to biased training data.

AI Ethics -  The study of moral issues and societal impacts of AI technology.
 

AI Training Data - Data used to train AI models to recognize patterns and make decisions.

Artificial General Intelligence (AGI) - General purpose systems with intelligence comparable to that of the human mind

Artificial Intelligence (AI)- The capacity of computers or other machines to exhibit or simulate intelligent behavior

Algorithm – A systematic and precise, step-by-step procedure (such as a recipe, a program, or a set of programs) for solving certain kinds of problems or accomplishing a task

Augmented Reality (AR) Technology that overlays digital information on the real world.  SLAM is often used to infer the world around you, which requires AI.

Automation - The conversion of a work process, a procedure, or equipment to automatic rather than human operation or control. 

Autonomous Systems - Systems capable of performing tasks without human intervention

Black Box - A device, system or object which can be viewed in terms of its inputs and outputs, without any knowledge of its internal workings

Big Data - Extremely large data sets that may be analyzed computationally to reveal patterns, trends, and associations

ChatGPT  An advanced language model developed by OpenAI, designed to generate human-like text based on user input. Utilizing natural language processing (NLP) and transformer architecture, ChatGPT can understand and respond to a wide range of topics, making it suitable for applications like customer support, content creation, and interactive conversational agents. It maintains context over conversations and can perform tasks such as answering questions, writing essays, and creating summaries. Although highly versatile, it occasionally produces incorrect responses, as it generates text based on patterns in its training data rather than factual knowledge.

Computer Vision - An AI labeling technique whereby computers process training images and videos in order to classify components from new imagery.

 

Convolutional Neural Networks (CNNs)  A type of neural network particularly effective for image recognition tasks.

Data Mining - The practice of examining large data sets in order to generate new information

Data Science - A multi-disciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from data Deep Learning - A class of machine learning algorithms that uses multiple layers to progressively extract higher level features from the raw input


Deep Learning -  A subset of Machine Learning that typically uses unstructured data to self solve complex problems.

Deepfake -  Manipulated photos and videos of people doing or saying something they never did, sometimes for

nefarious reasons, using deep learning

 

Edge AI - AI processing done on devices at the edge of the network rather than in a centralized data center.
 

Explainable A.I. -  A system that can detail factors that went into a prediction in a way that humans can interpret

Feature Extraction - The process of transforming raw data into numerical features for modeling.

Gemini  -  A series of advanced AI models developed by Google DeepMind. The latest version, Gemini 1.5 Pro, can handle large amounts of data and perform complex tasks across multiple modalities, including text, images, and video. For instance, it can analyze extensive transcripts, identify scenes from drawings, and solve challenging coding problems. This capability is enabled by its mixture-of-experts architecture, which activates specific neural network parts relevant to the task, improving efficiency and performance​ (MIT Technology Review)​​ (Google DeepMind)​.

Generative Adversarial Networks (GANs) - Neural networks that generate new data by pitting two models against each other.

Genetic Algorithm An optimization algorithm based on principles of genetics and natural selection

Image Processing - The analysis and manipulation of a digitized image, especially in order to improve its quality

Hyperparameter Tuning - The process of optimizing the parameters that control the learning process of an AI model.

Image Classification – Systems that sort images into predefined categories

Image Recognition  AI that can identify and process images to make decisions based on the visual input

Inference - Predicting an outcome based on a mathematical model given input conditions

Intelligent Agents -  Autonomous entities that observe and act upon an environment to achieve goals.

LLM - A large language model, such as ChatGPT, refers to a type of artificial intelligence that uses deep learning techniques to understand and generate human-like text. These models are trained on vast amounts of text data to learn patterns, understand language semantics, and generate coherent responses to queries or prompts. They can be used for various tasks like answering questions, generating text, translating languages, and more, depending on their training and programming.

Machine Learning - A subset of artificial intelligence that gives computer systems the ability to perform tasks without using explicit instructions.  Input data is usually structured.

Model – A mathematical or schematic description that describes the behavior of a system

 

Model Evaluation -  Assessing the performance of an AI model using various metrics and validation techniques

Natural Language Processing - A branch of artificial intelligence that deals with the interaction between computers and humans using human language syntax and semantics, rather than structured computer language

Neural Network – A supervised learning model inspired by the human brain made up units (neurons) connected in multiple layers

Overfitting and Underfitting - Overfitting occurs when a model is too complex, and underfitting occurs when a model is too simple to capture the data patterns.

Predictive Analytics - A variety of statistical techniques that analyze data to make predictions about future events.

Reinforcement Learning - The subfield of machine learning that learns from a series of reinforcements -- rewards or punishments

Recurrent Neural Networks (RNNs) - Neural networks designed for sequence data, like time series or natural language.

Simultaneous Localization and Mapping (SLAM) is a computational problem of constructing or updating a map of an unknown environment while simultaneously keeping track of an agent's location within it. It is commonly used in robotics and autonomous systems to navigate and map their surroundings without relying on pre-existing maps. SLAM algorithms combine data from various sensors, such as cameras, LiDAR, and IMUs (Inertial Measurement Units), to build a coherent map and determine the agent's position relative to it in real-time. This technology is crucial for applications like autonomous vehicles, drones, and augmented reality.

Supervised Learning - The subfield of machine learning that learns a function that maps an input to an output based on example input-output pairs

Training – The process by which supervised learning algorithms take input and tune a model to reflect the given data

 

Training Data  Data used to train AI models to recognize patterns and make decisions.

Transfer Learning - A machine learning technique that uses pre-trained models from one task to solve a different but related problem

Turing Test  A test for AI to determine if it can fool a person into believing they’re seeing or interacting with a real person.

Unsupervised Learning - The subfield of machine learning that learns structure

in the input without explicit feedback (like labels)

Virtual Reality (VR) -  Technology that creates a simulated environment for users to interact with

There are more terms out there, and we'll continue adding them as we learn ourselves!

bottom of page