AI Glossary

Demystify the world of AI with this curated glossary. Understand the terminology, explore the possibilities, and discover how AI impacts your life and the world.
AI Glossary
AI Explained

General AI

Abbreviation Term Definition Example Resources
AI Artificial Intelligence The ability of machines to exhibit intelligent behavior. Self-driving cars, virtual assistants, medical diagnosis systems. https://en.wikipedia.org/wiki/Artificial_intelligence
AGI Artificial General Intelligence A hypothetical form of AI able to understand or learn any intellectual task a human can. Currently non-existent, imagine an AI writing scientific papers and holding debates. https://en.wikipedia.org/wiki/Artificial_general_intelligence
ASI Artificial Superintelligence AI surpassing human intelligence in all aspects. Theoretical; potential dangers are debated by AI researchers. https://en.wikipedia.org/wiki/Superintelligence
ANI Artificial Narrow Intelligence AI designed to perform specific tasks well within a single domain. Chess-playing AI, spam filters, image recognition systems. https://en.wikipedia.org/wiki/Weak_artificial_intelligence
-- The Turing Test A test of a machine's ability to exhibit intelligent behavior indistinguishable from a human. Historically controversial, but an AI convincingly passing the Turing Test would be a major milestone. https://en.wikipedia.org/wiki/Turing_test
-- The Singularity A hypothetical point when technological growth, led by AI, spirals beyond control. Speculative, often in science fiction, but sparks debate about the long-term trajectory of AI. https://en.wikipedia.org/wiki/Technological_singularity

Machine Learning

Abbreviation Term Definition Example Resources
ML Machine Learning Field of AI where computers learn without explicit programming, through analyzing data and patterns. Spam filters, recommender systems, medical diagnosis, fraud detection https://developers.google.com/machine-learning/glossary
DL Deep Learning A subset of ML using complex neural networks loosely inspired by the human brain. Image classification, natural language processing, self-driving car perception https://www.deeplearning.ai/
NN Neural Network A mathematical model inspired by biological brains, forming the basis of many deep learning techniques. Image recognition, language translation, stock market prediction https://en.wikipedia.org/wiki/Artificial_neural_network
CNN Convolutional Neural Network A type of NN especially effective for image and video processing Facial recognition, object detection for self-driving cars, medical image analysis https://www.youtube.com/watch?v=FmpDIaiMIeA
RNN Recurrent Neural Network A type of NN designed to process sequential data, like text or time series. Machine translation, text summarization, stock price prediction https://en.wikipedia.org/wiki/Recurrent_neural_network
GAN Generative Adversarial Network A type of ML model where two networks compete, resulting in the ability to create realistic images, videos, or other data. Creating realistic synthetic images for art or training datasets, generating realistic product photos https://machinelearningmastery.com/what-are-generative-adversarial-networks-gans/
RL Reinforcement Learning An ML type focusing on an agent learning through trial-and-error interactions with an environment. Training robots to perform complex tasks, AI beating human champions in video games https://spinningup.openai.com/en/latest/
SL Supervised Learning An ML type where the model learns from a labeled dataset (input-output pairs). Facial recognition (learning from labeled images of faces), spam filters (learning from labeled examples of spam and non-spam emails) https://en.wikipedia.org/wiki/Supervised_learning
UL Unsupervised Learning An ML type where the model finds patterns in unlabeled data. Customer segmentation in marketing (identifying similar customer groups), anomaly detection in manufacturing https://en.wikipedia.org/wiki/Unsupervised_learning
SSL Semi-supervised Learning An ML type using a combination of small labeled data and a larger amount of unlabeled data. Medical image analysis (using some labeled samples, and many more unlabeled ones) improving speech recognition https://en.wikipedia.org/wiki/Semi-supervised_learning
-- Overfitting When a machine learning model learns the training data too closely, failing to generalize to new examples. An image classifier recognizing specific training images perfectly but poorly on unseen images https://aws.amazon.com/what-is/overfitting/
-- Transfer Learning Reusing knowledge gained from one ML task to improve performance on a related task. Pre-trained language model fine-tuned for specific industry tasks, saving time and data https://www.tensorflow.org/tutorials/images/transfer_learning
-- Probabilistic Reasoning A type of reasoning that takes into account uncertainty. Used in decision-making AI systems where outcomes are not certain, like weather prediction or medical diagnosis. https://en.wikipedia.org/wiki/Probabilistic_reasoning
-- Bayesian Reasoning A method of reasoning that uses probability theory to make inferences. Updating a spam filter after seeing new emails, robot navigation in unknown environments. https://en.wikipedia.org/wiki/Bayesian_inference
-- Genetic Algorithms A type of ML algorithm inspired by evolution, used to solve complex optimization problems Finding optimal product shipment routes, designing efficient neural network architectures. https://en.wikipedia.org/wiki/Genetic_algorithm

Natural Language Processing (NLP)

Abbreviation Term Definition Example Resources
NLP Natural Language Processing Computers understanding, interpreting, and manipulating human language. Chatbots, sentiment analysis, machine translation, text summarization https://nlp.stanford.edu/
NLU Natural Language Understanding An NLP subfield focusing on a computer's ability to read and understand text as a human would. Machines summarizing news articles, virtual assistants answering complex questions https://en.wikipedia.org/wiki/Natural_language_understanding
NLG Natural Language Generation An NLP subfield focusing on computers coherently producing human-like text. Automatic article writing, chatbots generating creative responses, composing personalized marketing emails https://en.wikipedia.org/wiki/Natural_language_generation
-- Sentiment Analysis Detecting the underlying emotions (positive, negative, neutral) within text. Businesses analyzing customer feedback on social media, tracking brand perception over time https://monkeylearn.com/sentiment-analysis/
-- Machine Translation Automatically translating text from one language to another. Google Translate, enabling cross-border communication, translating online product reviews https://en.wikipedia.org/wiki/Machine_translation
-- Text Summarization Condensing a long piece of text into a shorter version while retaining key information. News article summarizers, creating abstracts of research papers https://huggingface.co/transformers/task_summary.html
-- Named Entity Recognition (NER) Identifying and classifying named entities (people, organizations, locations) in text. Extracting important information from news articles or legal documents https://en.wikipedia.org/wiki/Named-entity_recognition
-- Topic Modeling Discovering abstract topics that occur within a collection of documents. Analyzing large datasets of customer reviews to identify common themes or issues. https://towardsdatascience.com/topic-modeling-and-latent-dirichlet-allocation-in-python-9bf156893c24
-- Speech Recognition The ability of computers to understand human speech. Virtual assistants like Siri or Alexa, automated transcription services. https://en.wikipedia.org/wiki/Speech_recognition

Computer Vision (CV)

Abbreviation Term Definition Example Resources
CV Computer Vision Enabling computers to see and interpret the visual world, similar to human vision. Object detection, image classification, facial recognition, autonomous robots https://opencv.org/
OCR Optical Character Recognition Converting images of typed, handwritten, or printed text into machine-readable text data. Scanning documents, extracting text from street signs in self-driving cars, reading handwritten forms https://en.wikipedia.org/wiki/Optical_character_recognition
-- Image Segmentation Partitioning an image into meaningful regions or objects. Medical image analysis (identifying tumors), separating foreground from background in photos for editing https://en.wikipedia.org/wiki/Image_segmentation
-- Object Detection Locating and identifying objects within an image or video. Self-driving cars detecting pedestrians and traffic signs, robots locating items to pick in a warehouse, facial recognition systems https://www.mathworks.com/discovery/object-detection.html
-- Image Classification Assigning labels to images based on their content. Sorting photos by type (landscape, portrait, etc.), medical image diagnosis (identifying different types of tumors) https://www.tensorflow.org/tutorials/images/classification
-- Pose Estimation Determining the position and orientation of a person or object in an image or video. Motion capture for animation in movies, tracking body movements in fitness applications https://www.youtube.com/watch?v=wEoyxE0GP2M
-- 3D Reconstruction Creating a 3D model of an object or a scene from images or videos. Applications in architecture (modeling buildings), augmented reality (placing virtual objects) https://en.wikipedia.org/wiki/3D_reconstruction

Hardware

Abbreviation Term Definition Example Resources
TPU Tensor Processing Unit ASICs (Application Specific Integrated Circuits) designed by Google, optimized for deep learning workloads. Accelerating the training of large AI models, particularly neural networks, often used in Google's data centers https://cloud.google.com/tpu
GPU Graphics Processing Unit Originally for graphics rendering, now widely used to accelerate ML tasks due to parallel processing power Training deep neural networks, processing images and videos, cryptocurrency mining https://www.nvidia.com/en-us/
LPU Large Language Model Processing Unit Emerging hardware specifically designed to handle the vast computational demands of large language models (LLMs). Aiming to process LLMs more efficiently than TPUs or GPUs, particularly for conversational AI tasks (Still in early development - Keep an eye on AI hardware news)
FPGA Field-Programmable Gate Array A type of integrated circuit that can be reconfigured after manufacturing, offering flexibility for specialized AI acceleration. Prototyping new AI chip designs, accelerating low-latency AI inference in edge devices https://en.wikipedia.org/wiki/Field-programmable_gate_array
CPU Central Processing Unit The general-purpose "brain" of a computer, capable of running various AI algorithms, though often with less efficiency than specialized hardware. Used for smaller AI models, AI tasks alongside general computing, or in devices where a dedicated accelerator is not cost-effective https://en.wikipedia.org/wiki/Central_processing_unit

Large-Scale Models (VLM, LLM)

Abbreviation Term Definition Example Resources
VLM Vision Language Model Deep learning models capable of processing both images and text. Generating image descriptions, answering questions about images, creating art based on text descriptions https://arxiv.org/abs/2307.05782 (Example research paper)
LLM Large Language Model Extremely powerful language models, trained on massive datasets of text, that can generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way ChatGPT, Google's Bard, Jurassic-1 Jumbo, other similar conversational AI tools https://openai.com/blog/chatgpt/
-- Transformer Architecture The neural network architecture underlying most LLMs and many VLMs, enabling them to process sequential data effectively. Self-attention mechanism in Transformers allows models to focus on relevant parts of text or images, improving their understanding https://www.youtube.com/watch?v=hYdO9CscNes (Explanatory Video)
-- Multimodal Models Models capable of processing and generating data of different modalities (e.g., text, image, audio) A model translating image descriptions into different languages, or generating a video clip based on a textual script https://arxiv.org/abs/2307.05782
-- Prompt Engineering The art of crafting prompts (instructions or input text) to get the best results from LLMs. Phrasing questions specifically, providing examples, or using keywords can influence the LLM's response style and accuracy https://en.wikipedia.org/wiki/Prompt_engineering

AI Techniques & Concepts

Term Definition Example Resources
Data Mining Extracting meaningful patterns and insights from large datasets. Businesses analyzing customer data to identify purchasing trends, scientific research finding patterns in genetic data https://en.wikipedia.org/wiki/Data_mining
Embeddings Mathematical representations of words or concepts, often used in NLP and recommendation systems. Representing a word like "cat" as a numerical vector, where similar words have similar vectors https://www.tensorflow.org/tutorials/text/word_embeddings
Knowledge Representation Techniques for how AI systems store and organize information. Semantic networks, ontologies used to represent relationships between concepts, knowledge graphs powering search engines https://en.wikipedia.org/wiki/Knowledge_representation_and_reasoning
Search Algorithms Methods used by AI systems to explore problem spaces and find solutions. A* search algorithm used for pathfinding in games and robotics, beam search for machine translation https://en.wikipedia.org/wiki/Search_algorithm
Expert Systems AI systems that emulate the decision-making ability of a human expert in a specific domain Medical diagnosis systems, financial risk assessment tools, systems guiding technicians through troubleshooting processes https://en.wikipedia.org/wiki/Expert_system
Robotics The integration of AI into robots and other physical devices. Autonomous warehouse robots, self-driving cars, robotic arms in manufacturing https://www.robotics.org/
Bias (in AI) Unfair or prejudicial outcomes resulting from AI decisions Facial recognition systems misidentifying certain ethnicities more often, algorithms biased against certain groups when making loan decisions https://www.ibm.com/blog/shedding-light-on-ai-bias-with-real-world-examples/
Explainable AI (XAI) Field focused on making the decisions of AI systems more transparent and understandable to humans. Tools for identifying which features of the input were most important for an AI model's decision, techniques for explaining why an image was classified a certain way https://www.darpa.mil/program/explainable-artificial-intelligence
AI Ethics Moral and social implications of AI, including fairness, transparency, accountability, and potential dangers. Ongoing discussion of safety protocols for self-driving cars, safeguards against biased AI in hiring, regulation of autonomous weapons https://en.wikipedia.org/wiki/Ethics_of_artificial_intelligence
Fuzzy Logic A type of logic that allows for uncertainty and partial truth. Controlling air conditioning systems, making self-driving car decisions in ambiguous situations. https://en.wikipedia.org/wiki/Fuzzy_logic
Pattern Recognition A fundamental ability for many kinds of AI tasks. Identifying objects in images, detecting fraudulent transactions, classifying types of music. https://en.wikipedia.org/wiki/Pattern_recognition
Recommender Systems Systems recommending products or services to users based on their patterns and data. Netflix suggesting movies you might like, Amazon recommending products to buy. https://en.wikipedia.org/wiki/Recommender_system
Symbolic AI A classic approach to AI, using symbols and rules to represent knowledge, contrasting with data-driven methods. Expert systems for medical diagnosis (in the past), AI systems designed to play chess. https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence

This glossary can be useful for:

  • Anyone who wants to learn more about AI.
  • AI researchers and professionals who want a quick reference.
  • Educators and students who are teaching or learning about AI.

Notes:

  • AI is a rapidly evolving field, so new terms and definitions may be created over time.
  • There may be multiple definitions for the same term.
  • It is important to consider the context when using AI terminology.

I hope you find this glossary helpful!

AI Pill

Take AI 💊 Deep Dive Into The Coming Wave.

AI Pill

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to AI Pill.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.