This glossary provides a foundational understanding of key terms and concepts in the rapidly evolving field of artificial intelligence, helping you navigate discussions and developments in AI more effectively.
Algorithm:
- Definition: A step-by-step set of rules or instructions designed to perform a specific task or solve a problem.
- Example: The PageRank algorithm used by Google to rank web pages in search results.
Artificial General Intelligence (AGI):
- Definition: A theoretical form of AI that possesses the ability to perform any intellectual task that a human can do, exhibiting general cognitive abilities.
- Example: AGI would be capable of learning and adapting to new tasks and environments without human intervention, unlike current AI systems which are task-specific.
Artificial Intelligence (AI):
- Definition: A branch of computer science that aims to create machines capable of performing tasks that typically require human intelligence, such as learning, reasoning, problem-solving, and understanding natural language.
- Example: AI is used in applications like virtual assistants (e.g., Siri, Alexa) and recommendation systems (e.g., Netflix, Amazon).
Artificial Neural Network (ANN):
- Definition: A computing system modeled after the human brain’s neural networks, designed to recognize patterns and solve problems through layers of interconnected nodes.
- Example: ANNs are used in image recognition systems to identify objects in photographs.
Automated Decision-Making:
- Definition: The process by which AI systems make decisions without human intervention, often based on pre-defined rules or learned patterns from data.
- Example: Credit scoring systems that automatically approve or reject loan applications based on an applicant’s financial data.
Backpropagation:
- Definition: A method used in training neural networks, where the model’s error is calculated and propagated backward through the network to adjust the weights, improving accuracy.
- Example: Backpropagation is used to refine a neural network’s predictions, such as identifying objects in images more accurately.
Bias in AI:
- Definition: The presence of systematic errors or prejudices in AI algorithms that lead to unfair or inaccurate outcomes, often due to biased training data.
- Example: An AI model trained on biased hiring data might unfairly disadvantage certain groups of candidates.
Big Data:
- Definition: Extremely large datasets that are too complex for traditional data processing methods, often requiring advanced analytics and AI techniques to extract insights.
- Example: Social media platforms generate big data from millions of users posting, liking, and sharing content every day.
Computer Vision:
- Definition: A field of AI that enables machines to interpret and make decisions based on visual data, such as images or videos.
- Example: Autonomous vehicles use computer vision to detect and respond to objects on the road.
Convolutional Neural Network (CNN):
- Definition: A type of deep learning neural network particularly well-suited for processing grid-like data, such as images, using layers that convolve over the input data.
- Example: CNNs are used in computer vision tasks like detecting faces in security cameras.
Data Mining:
- Definition: The process of discovering patterns, correlations, and insights from large datasets using statistical and machine learning techniques.
- Example: Retailers use data mining to analyze purchase histories and predict future buying trends.
Deep Learning:
- Definition: A type of machine learning that uses neural networks with many layers (hence “deep”) to model complex patterns in data.
- Example: Deep learning is used in image recognition software, such as facial recognition in social media platforms.
Deepfakes:
- Definition: AI-generated synthetic media where an existing image or video is altered to create fake but realistic representations of people or events.
- Example: Deepfake videos where a person’s face is superimposed onto another body, often used to create misleading or humorous content.
Explainable AI (XAI):
- Definition: AI systems designed to provide human-understandable explanations for their decisions and actions, addressing the need for transparency and trust in AI applications.
- Example: A credit scoring AI providing reasons for approving or denying a loan application.
Feature Engineering:
- Definition: The process of selecting, modifying, or creating features (input variables) that improve the performance of a machine learning model.
- Example: Transforming raw data on dates into features like day of the week, month, or holiday indicator for a model predicting sales.
Fuzzy Logic:
- Definition: A form of logic used in AI that allows for reasoning with degrees of truth, rather than the traditional binary true/false logic.
- Example: Fuzzy logic is used in smart home systems to adjust heating based on various degrees of comfort rather than fixed temperature settings.
Generative Adversarial Network (GAN):
- Definition: A type of neural network where two networks (a generator and a discriminator) compete with each other to create data that is indistinguishable from real data.
- Example: GANs are used to generate realistic images, such as creating human faces that don’t exist.
Generative AI:
- Definition: A type of AI that creates new content, such as images, text, or music, by learning patterns from existing data.
- Example: AI systems like GPT-3 generate human-like text for chatbots, content creation, and more.
Hallucinations:
- Definition: In AI, particularly in language models, hallucinations refer to the generation of outputs that are plausible-sounding but factually incorrect or nonsensical.
- Example: A language model incorrectly generating historical facts that are not accurate or true.
Hyperparameters:
- Definition: Parameters in a machine learning model that are set before the learning process begins and control the training process, such as learning rate and the number of layers in a neural network.
- Example: The learning rate in a gradient descent algorithm is a hyperparameter that determines the size of steps taken towards the minimum.
Large Language Models (LLMs):
- Definition: AI models trained on vast amounts of text data to understand, generate, and manipulate human language at a sophisticated level.
- Example: GPT-3 and BERT are large language models used for tasks like translation, summarization, and conversational AI.
Machine Learning (ML):
- Definition: A subset of AI that involves training algorithms on data so they can make predictions or decisions without being explicitly programmed for specific tasks.
- Example: Spam filters in email use machine learning to identify and filter out unwanted messages.
Natural Language Generation (NLG):
- Definition: The process of generating human-like text from structured data using AI techniques, often used in applications like automated report writing and chatbots.
- Example: An AI system generating news articles based on sports event data.
Natural Language Processing (NLP):
- Definition: A field of AI focused on the interaction between computers and humans through natural language, enabling machines to understand, interpret, and respond to human language.
- Example: NLP powers chatbots, language translation apps, and sentiment analysis tools.
Neural Network:
- Definition: A computational model inspired by the human brain, consisting of layers of interconnected nodes (neurons) that process information and learn from data.
- Example: Neural networks are used in voice recognition systems to transcribe speech to text.
Overfitting:
- Definition: A modeling error that occurs when a machine learning model learns the details and noise in the training data to an extent that it negatively impacts its performance on new data.
- Example: A model trained on a small, specific dataset may perform well on that data but fail to generalize to unseen data.
Reinforcement Learning:
- Definition: A type of machine learning where an agent learns by interacting with its environment, receiving rewards or penalties based on its actions.
- Example: Training a robot to navigate a maze by rewarding it for reaching the end.
Supervised Learning:
- Definition: A type of machine learning where the model is trained on a labeled dataset, meaning the correct output is provided during training.
- Example: Predicting house prices based on historical data where each house’s price is labeled.
Tokenization:
- Definition: The process of converting a piece of text into smaller units called tokens, which can be words, phrases, or even characters, used in NLP tasks.
- Example: Tokenizing the sentence “AI is fascinating” into [“AI”, “is”, “fascinating”].
Transfer Learning:
- Definition: A machine learning technique where a model developed for one task is reused as the starting point for a model on a second, related task.
- Example: A model trained to recognize objects in photos can be fine-tuned to detect specific types of objects, like medical anomalies in X-rays.
Turing Test:
- Definition: A test proposed by Alan Turing to determine whether a machine can exhibit intelligent behavior indistinguishable from that of a human.
- Example: A chatbot passing the Turing Test would be able to converse with a human without the human realizing they are speaking to a machine.
Unsupervised Learning:
- Definition: A type of machine learning where the model is trained on data without explicit labels, allowing the algorithm to discover hidden patterns or relationships.
- Example: Clustering customers into segments based on purchasing behavior without predefined categories.