AI Topics
Short, factual explainers for the artificial-intelligence terms people search most — from ChatGPT and large language models to AGI. Search or browse the 607 entries below.
No topics match your search. Try a different term.
Models & Products
- What Is ChatGPT? ChatGPT is OpenAI’s conversational AI chatbot, launched in November 2022, built on large language models and trained to follow instructions.
- What Is GPT-4? GPT-4 is OpenAI’s large multimodal language model, released in 2023, capable of understanding text and images and powering advanced versions of ChatGPT.
- What Is GPT-4o? GPT-4o is OpenAI’s “omni” model that handles text, images, and audio in a single system, released in 2024 for faster, natural interaction.
- What Is OpenAI? OpenAI is the AI research company behind ChatGPT, GPT-4, DALL·E, and Sora, known for popularizing generative AI.
- What Is Google Gemini? Gemini is Google DeepMind’s family of multimodal AI models and the assistant that replaced Bard.
- What Is Claude (Anthropic)? Claude is a family of AI assistants built by Anthropic, known for long context windows and a focus on safety.
- What Is Llama (Meta AI)? Llama is Meta’s family of open-weight large language models that developers can download, run, and fine-tune themselves.
- What Is Mistral AI? Mistral AI is a European company known for efficient, high-performing open-weight language models.
- What Is DeepSeek? DeepSeek is a Chinese AI lab whose efficient open models, including a reasoning model, drew global attention in early 2025.
- What Is Grok (xAI)? Grok is the conversational AI model built by xAI, Elon Musk’s AI company, integrated with the X platform.
- What Is Microsoft Copilot? Microsoft Copilot is an AI assistant built into Windows, Microsoft 365, and Edge, powered by large language models.
- What Is GitHub Copilot? GitHub Copilot is an AI pair-programmer that suggests code and whole functions inside the editor.
- What Is Perplexity AI? Perplexity is an AI-powered answer engine that responds to questions with cited, sourced summaries.
- What Is Midjourney? Midjourney is a popular AI image generator that creates detailed pictures from text prompts.
- What Is Stable Diffusion? Stable Diffusion is an open-source text-to-image model that can run on consumer hardware.
- What Is DALL·E? DALL·E is OpenAI’s text-to-image system that generates original images from written descriptions.
- What Is Sora (Text-to-Video)? Sora is OpenAI’s text-to-video model that generates short, realistic video clips from prompts.
- What Is Runway? Runway is a company building AI video-generation and editing tools used in film and creative work.
- What Is ElevenLabs? ElevenLabs is an AI voice company known for realistic text-to-speech and voice cloning.
- What Is Suno AI? Suno is an AI tool that generates complete songs, including vocals and instruments, from text prompts.
- What Is NotebookLM? NotebookLM is Google’s AI research assistant that answers questions grounded in your own documents.
- What Is Character.AI? Character.AI is a platform for chatting with AI characters and personas created by users.
- What Is Hugging Face? Hugging Face is a hub for open-source AI models, datasets, and tools used by the ML community.
- What Are AI Reasoning Models? Reasoning models are LLMs that “think” step by step before answering, improving performance on hard problems.
Core Concepts
- What Is a Large Language Model (LLM)? A large language model is an AI trained on vast text to predict and generate language, powering tools like ChatGPT.
- What Is Generative AI? Generative AI creates new content — text, images, audio, video, or code — rather than only classifying data.
- What Is a Transformer in AI? The transformer is the neural network architecture, introduced in 2017, behind modern language and multimodal models.
- What Is the Attention Mechanism? Attention lets a model weigh which parts of the input matter most for each part of the output.
- What Is Prompt Engineering? Prompt engineering is the practice of writing effective instructions to get better results from AI models.
- What Is Retrieval-Augmented Generation (RAG)? RAG combines a language model with a search step so answers are grounded in specific, up-to-date documents.
- What Is Fine-Tuning in AI? Fine-tuning adapts a pretrained model to a specific task or style by training it further on focused data.
- What Is an AI Hallucination? An AI hallucination is when a model states false information confidently and fluently as if it were true.
- What Are Tokens in AI? Tokens are the small chunks of text — words or word pieces — that language models read and generate.
- What Is Tokenization? Tokenization is the step that splits raw text into tokens a model can process.
- What Are Embeddings in AI? Embeddings turn words, images, or data into vectors so that similar items sit close together in space.
- What Is a Vector Database? A vector database stores embeddings and finds the most similar items quickly, powering AI search and RAG.
- What Is a Context Window? The context window is how much text a model can consider at once, measured in tokens.
- What Is a Diffusion Model? Diffusion models generate images by gradually turning random noise into a coherent picture.
- What Is a GAN (Generative Adversarial Network)? A GAN pits two neural networks against each other to generate realistic images and other data.
- What Is a Neural Network? A neural network is a system of interconnected units that learns patterns from data by adjusting weights.
- What Is Deep Learning? Deep learning uses many-layered neural networks to learn features directly from raw data.
- What Is Machine Learning? Machine learning is the field of building systems that learn patterns from data instead of following fixed rules.
- What Is Reinforcement Learning? Reinforcement learning trains agents to make decisions by rewarding good outcomes and penalizing bad ones.
- What Is RLHF (Reinforcement Learning from Human Feedback)? RLHF fine-tunes AI models using human preferences, helping make chatbots helpful and better aligned.
- What Is Multimodal AI? Multimodal AI understands and generates across several types of data — text, images, audio, and video.
- What Are AI Agents? AI agents are systems that plan and take multi-step actions toward a goal, using tools with some autonomy.
- What Is Agentic AI? Agentic AI describes systems that act autonomously over multiple steps to achieve goals, not just respond.
- What Is a Foundation Model? A foundation model is a large model trained on broad data that can be adapted to many downstream tasks.
- What Are Parameters in an AI Model? Parameters are the learned values inside a model; their number is a rough measure of model size.
- What Is Inference in AI? Inference is the stage where a trained model is used to make predictions or generate output.
- What Is Training Data? Training data is the material a model learns from; its quality and content shape the model’s behavior.
- What Is Overfitting? Overfitting is when a model memorizes its training data and fails to generalize to new examples.
- What Is Backpropagation? Backpropagation is the algorithm that trains neural networks by adjusting weights to reduce error.
- Supervised vs. Unsupervised Learning Supervised learning trains on labeled examples; unsupervised learning finds structure in unlabeled data.
- What Are Zero-Shot and Few-Shot Learning? Zero-shot and few-shot learning let a model handle new tasks with no or only a few examples in the prompt.
- What Is Chain-of-Thought Prompting? Chain-of-thought prompting asks a model to reason step by step, improving results on complex problems.
- What Is Temperature in AI Text Generation? Temperature is a setting that controls how random or focused a model’s generated text is.
- What Is Model Quantization? Quantization shrinks a model by storing its numbers at lower precision, so it runs faster and cheaper.
- What Is a Mixture of Experts (MoE)? Mixture of experts activates only part of a large model per input, cutting cost while keeping capacity.
- What Is Knowledge Distillation? Distillation trains a small “student” model to mimic a larger “teacher,” keeping much of its ability.
Tools & Applications
- What Is an AI Chatbot? An AI chatbot is a program that holds natural conversations, now often powered by large language models.
- What Is AI Art? AI art is imagery generated by models from text prompts, popularized by tools like Midjourney and DALL·E.
- What Is Text-to-Image AI? Text-to-image AI turns a written description into an original picture using generative models.
- What Is Text-to-Video AI? Text-to-video AI generates short video clips from written prompts, as with OpenAI’s Sora.
- What Is an AI Coding Assistant? AI coding assistants suggest, complete, and explain code, speeding up software development.
- What Are AI Writing Tools? AI writing tools help draft, edit, and rewrite text, from emails to articles, using language models.
- What Is a Deepfake? A deepfake is synthetic media that convincingly shows a real person saying or doing something they never did.
- What Is AI Voice Cloning? Voice cloning uses AI to recreate a specific person’s voice from a short audio sample.
- What Is an AI Content Detector? AI detectors try to judge whether text or media was generated by AI, though their reliability is limited.
- How Is AI Used in Healthcare? AI supports medical imaging, diagnosis, and research, working alongside clinicians rather than replacing them.
- How Is AI Used in Education? AI enables adaptive tutoring and instant feedback, supplementing teachers rather than replacing them.
- How Do Self-Driving Cars Use AI? Autonomous vehicles use AI and computer vision to perceive their surroundings and make driving decisions.
- What Are Recommendation Systems? Recommendation systems predict what a user will like, powering feeds on streaming, shopping, and social apps.
- What Is Computer Vision? Computer vision is the field of AI that enables machines to interpret images and video.
- What Is Natural Language Processing (NLP)? NLP is the field of AI focused on understanding and generating human language.
- What Is Speech Recognition? Speech recognition converts spoken words into text, powering dictation and voice assistants.
- What Is Sentiment Analysis? Sentiment analysis uses AI to detect the emotion or opinion expressed in text.
- What Are AI Search Engines? AI search engines answer questions with synthesized, cited summaries instead of just links.
- How Does AI Accelerate Scientific Research? AI narrows vast search spaces in fields like protein folding and materials discovery, speeding up science.
Safety, Ethics & Future
- What Is Artificial General Intelligence (AGI)? AGI is a hypothetical AI with human-level general ability across tasks, unlike today’s narrow systems.
- What Is Superintelligence? Superintelligence is a hypothetical AI far surpassing human intelligence in virtually every domain.
- What Is AI Safety? AI safety is the field working to ensure AI systems behave reliably and do not cause unintended harm.
- What Is AI Alignment? AI alignment is the challenge of making AI systems pursue the goals and values their designers intend.
- What Is AI Bias? AI bias is when a model reproduces or amplifies unfair patterns learned from its training data.
- What Is AI Ethics? AI ethics examines how to build and use AI responsibly — addressing bias, privacy, transparency, and harm.
- What Is AI Regulation? AI regulation is the set of laws and rules governing how AI can be built and used, such as the EU AI Act.
- What Is the EU AI Act? The EU AI Act is Europe’s risk-based law regulating artificial intelligence, adopted in 2024.
- How Will AI Affect Jobs? AI can automate routine tasks, reshaping jobs — displacing some roles while creating and changing others.
- What Is Existential Risk from AI? Existential risk refers to the debated possibility that highly advanced AI could threaten humanity’s future.
- What Is the AI Control Problem? The control problem asks how humans can keep highly capable AI systems doing what we want.
- What Is Explainable AI (XAI)? Explainable AI aims to make model decisions understandable, so people can trust and audit them.
- AI and Misinformation Generative AI lowers the cost of producing convincing false text, images, and video at scale.
- AI and Surveillance AI enables monitoring at new scale through facial recognition and behavioral analysis, raising privacy concerns.
- What Is Open-Source AI? Open-source (or open-weight) AI makes model weights available to download, run, and modify freely.
- AI and Copyright AI raises unresolved questions about training on copyrighted work and who owns AI-generated output.
History & People
- The History of Artificial Intelligence AI’s history runs from 1950s foundations through symbolic AI, expert systems, machine learning, and today’s LLMs.
- What Was the Dartmouth Conference? The 1956 Dartmouth workshop named artificial intelligence and founded it as a field of research.
- Who Was John McCarthy? John McCarthy coined “artificial intelligence,” organized the 1956 Dartmouth conference, and created LISP.
- Who Was Alan Turing? Alan Turing was a pioneer of computing whose 1950 work framed the question of machine intelligence.
- What Is the Turing Test? The Turing test asks whether a machine can converse indistinguishably from a human.
- What Is the LISP Programming Language? LISP, created by John McCarthy in 1958, was the dominant language of early AI research.
- What Was an AI Winter? An AI winter is a period when funding and interest in AI collapsed after results fell short of hype.
- What Were Expert Systems? Expert systems were 1980s AI programs that encoded specialists’ knowledge as large sets of if–then rules.
- What Is Symbolic AI? Symbolic AI represents knowledge as symbols and reasons over them with explicit logical rules.
- What Was AlphaGo? AlphaGo was DeepMind’s program that beat a top human Go player in 2016, a landmark for AI.
- What Was Deep Blue? Deep Blue was IBM’s chess computer that beat world champion Garry Kasparov in 1997.
- What Is ImageNet? ImageNet is a large labeled image dataset whose 2012 benchmark sparked the deep-learning revolution.
Machine Learning & Training
- What Is Gradient Descent? Gradient descent is an optimization algorithm that iteratively adjusts model parameters to minimize a loss function.
- What Is Stochastic Gradient Descent? Stochastic gradient descent updates model parameters using one or a few training examples at a time rather than the full dataset.
- What Is the Adam Optimizer? Adam is an adaptive optimization algorithm that combines momentum and per-parameter learning rates to train neural networks.
- What Is Learning Rate? The learning rate is a hyperparameter that controls how large a step a model takes when updating its parameters during training.
- What Is a Loss Function? A loss function measures how far a model's predictions are from the true values, guiding optimization during training.
- What Is Cross-Entropy Loss? Cross-entropy loss measures the difference between predicted probability distributions and true labels in classification tasks.
- What Is Mean Squared Error? Mean squared error is a loss metric that averages the squared differences between predicted and actual values.
- What Is Regularization? Regularization is a set of techniques that discourage overly complex models to reduce overfitting and improve generalization.
- What Is L1 Regularization? L1 regularization adds a penalty proportional to the absolute value of model weights, often producing sparse models.
- What Is L2 Regularization? L2 regularization penalizes large model weights by adding the squared magnitude of weights to the loss function.
- What Is Dropout? Dropout is a regularization method that randomly deactivates neurons during training to reduce overfitting.
- What Is Batch Normalization? Batch normalization normalizes layer inputs across a mini-batch to stabilize and speed up neural network training.
- What Is Layer Normalization? Layer normalization normalizes the activations within each individual example across features rather than across a batch.
- What Is an Activation Function? An activation function introduces non-linearity into a neural network, enabling it to learn complex patterns.
- What Is ReLU? ReLU is an activation function that outputs the input if positive and zero otherwise, widely used in deep networks.
- What Is the Sigmoid Function? The sigmoid function maps any real number to a value between 0 and 1, often used to output probabilities.
- What Is Softmax? Softmax converts a vector of numbers into a probability distribution that sums to one, used for multi-class classification.
- What Is the Tanh Activation Function? The tanh activation function maps inputs to values between -1 and 1, offering a zero-centered alternative to sigmoid.
- What Is an Epoch? An epoch is one complete pass through the entire training dataset during the training of a machine learning model.
- What Is Batch Size? Batch size is the number of training examples processed together before the model's parameters are updated.
- What Is a Mini-Batch? A mini-batch is a small subset of training data used to compute a gradient update, balancing speed and stability.
- What Is Gradient Clipping? Gradient clipping limits the size of gradients during training to prevent unstable updates and exploding gradients.
- What Is the Vanishing Gradient Problem? The vanishing gradient problem occurs when gradients become too small for effective learning in deep networks.
- What Is the Exploding Gradient Problem? The exploding gradient problem happens when gradients grow too large during training, destabilizing the model.
- What Is Weight Initialization? Weight initialization is the process of setting a neural network's initial parameter values before training begins.
- What Is Momentum Optimization? Momentum optimization accelerates gradient descent by accumulating past gradients to smooth and speed up updates.
- What Is a Learning Rate Schedule? A learning rate schedule adjusts the learning rate over the course of training to improve convergence.
- What Is Early Stopping? Early stopping halts training when validation performance stops improving, helping to prevent overfitting.
- What Is Cross-Validation? Cross-validation evaluates a model by training and testing it on different splits of the data to estimate performance.
- What Is K-Fold Cross-Validation? K-fold cross-validation splits data into k parts, training on some and testing on the rest across multiple rounds.
- What Is a Train-Test Split? A train-test split divides data into separate sets for training a model and evaluating its performance.
- What Is a Validation Set? A validation set is data used to tune hyperparameters and monitor model performance during training.
- What Is Underfitting? Underfitting occurs when a model is too simple to capture the underlying patterns in the data.
- What Is the Bias-Variance Tradeoff? The bias-variance tradeoff describes the balance between a model's simplicity and its sensitivity to training data.
- What Is Feature Engineering? Feature engineering is the process of creating and transforming input variables to improve model performance.
- What Is Feature Selection? Feature selection identifies the most relevant input variables to improve model accuracy and reduce complexity.
- What Is Feature Scaling? Feature scaling adjusts the range of input features so they contribute comparably during model training.
- What Is Normalization? Normalization rescales data to a fixed range, commonly between 0 and 1, to make features comparable.
- What Is Standardization? Standardization rescales features to have zero mean and unit variance for consistent model training.
- What Is One-Hot Encoding? One-hot encoding converts categorical variables into binary vectors so models can process them numerically.
- What Is Dimensionality Reduction? Dimensionality reduction lowers the number of input features while preserving important information in the data.
- What Is Principal Component Analysis? Principal component analysis reduces dimensionality by projecting data onto directions of greatest variance.
- What Is t-SNE? t-SNE is a technique for visualizing high-dimensional data by mapping it into two or three dimensions.
- What Is UMAP? UMAP is a dimensionality reduction method that preserves data structure for visualization and analysis.
- What Is Unsupervised Learning? Unsupervised learning finds patterns and structure in data without using labeled examples.
- What Is Semi-Supervised Learning? Semi-supervised learning combines a small amount of labeled data with a large amount of unlabeled data.
- What Is Self-Supervised Learning? Self-supervised learning creates training signals from unlabeled data by predicting parts of the input itself.
- What Is Transfer Learning? Transfer learning reuses knowledge from a model trained on one task to improve learning on a related task.
- What Is Ensemble Learning? Ensemble learning combines multiple models to produce more accurate and robust predictions than any single model.
- What Is Bagging? Bagging trains multiple models on random data samples and averages their predictions to reduce variance.
- What Is Boosting? Boosting builds models sequentially, with each one correcting the errors of the previous, to form a strong learner.
- What Is Gradient Boosting? Gradient boosting builds an ensemble of models sequentially, each fitting the residual errors of the previous ones.
- What Is XGBoost? XGBoost is a fast, scalable implementation of gradient boosting widely used for structured data tasks.
- What Is LightGBM? LightGBM is a gradient boosting framework designed for speed and efficiency on large datasets.
- What Is a Random Forest? A random forest is an ensemble of decision trees whose predictions are combined for improved accuracy.
- What Is a Decision Tree? A decision tree is a model that makes predictions by splitting data through a series of yes-or-no questions.
- What Is a Support Vector Machine? A support vector machine classifies data by finding the boundary that best separates different classes.
- What Is K-Nearest Neighbors? K-nearest neighbors classifies or predicts based on the most similar examples in the training data.
- What Is Naive Bayes? Naive Bayes is a probabilistic classifier based on Bayes theorem assuming features are independent.
- What Is Logistic Regression? Logistic regression is a statistical model used to predict the probability of a categorical outcome.
- What Is Linear Regression? Linear regression models the relationship between inputs and a continuous output using a straight-line fit.
- What Is Clustering? Clustering groups similar data points together without using labels, revealing structure in the data.
- What Is K-Means Clustering? K-means clustering partitions data into k groups by minimizing the distance between points and cluster centers.
- What Is Hierarchical Clustering? Hierarchical clustering builds a tree of nested clusters by merging or splitting groups of data points.
- What Is DBSCAN? DBSCAN is a clustering algorithm that groups dense regions of points and labels sparse points as noise.
- What Is a Gaussian Mixture Model? A Gaussian mixture model represents data as a combination of several Gaussian distributions for clustering.
- What Is Anomaly Detection? Anomaly detection identifies rare data points that differ significantly from the expected pattern.
- What Is a Confusion Matrix? A confusion matrix summarizes classification results by counting correct and incorrect predictions per class.
- What Are Precision and Recall? Precision and recall are metrics that measure the accuracy and completeness of a classifier's positive predictions.
- What Is the F1 Score? The F1 score is the harmonic mean of precision and recall, balancing both into a single metric.
- What Is an ROC Curve? An ROC curve plots a classifier's true positive rate against its false positive rate across thresholds.
- What Is AUC? AUC is the area under the ROC curve, summarizing a classifier's ability to distinguish between classes.
- What Is Accuracy in Machine Learning? Accuracy measures the fraction of predictions a model gets correct out of all predictions made.
- What Is Hyperparameter Tuning? Hyperparameter tuning is the process of searching for the settings that make a model perform best.
- What Is Grid Search? Grid search tunes hyperparameters by exhaustively testing every combination in a predefined set of values.
- What Is Random Search? Random search tunes hyperparameters by sampling random combinations rather than testing every option.
- What Is Bayesian Optimization? Bayesian optimization efficiently tunes hyperparameters by building a probabilistic model of performance.
- What Is Data Augmentation? Data augmentation expands training data by creating modified copies of existing examples to improve models.
- What Is Class Imbalance? Class imbalance occurs when some categories appear far more often than others in a dataset.
- What Is a Label in Machine Learning? A label is the correct output or target value assigned to a training example in supervised learning.
- What Is Ground Truth? Ground truth is the verified correct data used to train and evaluate machine learning models.
- What Is Model Evaluation? Model evaluation measures how well a trained model performs using metrics and held-out data.
- What Is Generalization? Generalization is a model's ability to perform well on new, unseen data rather than just training data.
- What Is the Curse of Dimensionality? The curse of dimensionality describes how data becomes sparse and harder to model as features increase.
- What Is a Markov Decision Process? A Markov decision process is a mathematical framework for modeling decision-making with states, actions, and rewards.
- What Is Q-Learning? Q-learning is a reinforcement learning algorithm that learns the value of actions to guide decision-making.
- What Is a Policy Gradient? Policy gradient methods optimize an agent's decision policy directly by following the gradient of expected reward.
- What Is Actor-Critic? Actor-critic methods combine a policy learner and a value estimator to improve reinforcement learning stability.
- What Is a Reward Function? A reward function defines the goal in reinforcement learning by assigning value to an agent's actions.
- What Is Exploration vs Exploitation? Exploration vs exploitation is the tradeoff between trying new actions and using known rewarding ones.
- What Is a Monte Carlo Method? Monte Carlo methods use repeated random sampling to estimate values and solve problems numerically.
- What Is the Bellman Equation? The Bellman equation expresses the value of a state in terms of immediate reward and future state values.
- What Is Contrastive Learning? Contrastive learning trains models to pull similar examples together and push dissimilar ones apart.
- What Is Curriculum Learning? Curriculum learning trains models on easier examples first before gradually introducing harder ones.
- What Is Active Learning? Active learning lets a model select the most informative examples for labeling to learn efficiently.
- What Is Online Learning? Online learning updates a model incrementally as new data arrives rather than training on a fixed dataset.
- What Is Meta-Learning? Meta-learning, or learning to learn, helps models adapt quickly to new tasks from limited experience.
- What Is Zero-Shot Learning? Zero-shot learning enables a model to handle tasks or classes it was never explicitly trained on.
- What Is One-Shot Learning? One-shot learning enables a model to recognize a new class from just a single labeled example.
- What Is an Autoencoder? An autoencoder is a neural network that learns to compress data and reconstruct it from a compact representation.
Language & LLMs
- What Is Byte-Pair Encoding? Byte-pair encoding is a subword tokenization method that merges frequent character pairs to build a vocabulary for language models.
- What Is Subword Tokenization? Subword tokenization splits words into smaller units so language models can handle rare words and large vocabularies efficiently.
- What Is a Word Embedding? A word embedding is a numerical vector representing a word's meaning, placing similar words near each other in vector space.
- What Is word2vec? word2vec is a technique that learns word embeddings from text by predicting words from their surrounding context.
- What Are GloVe Embeddings? GloVe embeddings are word vectors learned from global word co-occurrence statistics across a large text corpus.
- What Is a Contextual Embedding? A contextual embedding is a word representation that changes based on the surrounding words in a sentence.
- What Is Positional Encoding? Positional encoding adds order information to token embeddings so transformers know the sequence position of each word.
- What Is Self-Attention? Self-attention lets a model weigh the importance of other tokens in a sequence when representing each token.
- What Is Multi-Head Attention? Multi-head attention runs several attention operations in parallel so a model can capture different relationships at once.
- What Is Cross-Attention? Cross-attention lets one sequence attend to another, connecting an encoder's output to a decoder in transformers.
- What Is an Encoder-Decoder Architecture? An encoder-decoder architecture reads an input sequence with one component and generates an output sequence with another.
- What Is a Transformer Encoder? A transformer encoder maps an input sequence into contextual representations using stacked self-attention layers.
- What Is a Transformer Decoder? A transformer decoder generates output tokens one at a time using masked self-attention over previous tokens.
- What Is BERT? BERT is a transformer encoder model that learns bidirectional context using masked language modeling.
- What Is the GPT Model Family? The GPT model family is a series of autoregressive transformer decoders that generate text by predicting the next token.
- What Is GPT-3? GPT-3 is a large autoregressive language model by OpenAI known for strong few-shot and in-context learning abilities.
- What Is GPT-3.5? GPT-3.5 is a family of OpenAI language models refined with instruction tuning and used in early versions of ChatGPT.
- What Is T5? T5 is a text-to-text transformer that frames every NLP task as converting an input string into an output string.
- What Is RoBERTa? RoBERTa is an optimized version of BERT trained longer on more data with tuned settings for better performance.
- What Is XLNet? XLNet is a language model that learns bidirectional context using permutation-based autoregressive training.
- What Is ELECTRA? ELECTRA is a pretraining method that trains a model to detect replaced tokens rather than predict masked ones.
- What Is Sequence to Sequence? Sequence to sequence models transform one sequence into another, such as translating a sentence into another language.
- What Is a Recurrent Neural Network? A recurrent neural network processes sequences step by step, passing a hidden state forward to capture context.
- What Is an LSTM? An LSTM is a recurrent network with gating mechanisms that help it learn long-range dependencies in sequences.
- What Is a GRU? A GRU is a simplified gated recurrent network that captures sequence dependencies with fewer parameters than an LSTM.
- What Is a Bidirectional RNN? A bidirectional RNN reads a sequence in both directions to combine past and future context for each position.
- What Is a Language Model? A language model assigns probabilities to sequences of words and predicts likely next tokens in text.
- What Is Masked Language Modeling? Masked language modeling trains a model to predict hidden words in a sentence using surrounding context.
- What Is Next-Token Prediction? Next-token prediction trains a model to guess the following token in a sequence, the core objective of GPT models.
- What Is an Autoregressive Model? An autoregressive model generates each output token conditioned on the tokens it has already produced.
- What Is a Causal Language Model? A causal language model predicts each token using only preceding tokens, preventing it from seeing future context.
- What Is Perplexity as a Metric? Perplexity measures how well a language model predicts text; lower perplexity indicates better predictions.
- What Is Beam Search? Beam search is a decoding strategy that keeps several candidate sequences at each step to find a high-probability output.
- What Is Greedy Decoding? Greedy decoding generates text by always choosing the single most probable next token at each step.
- What Is Top-k Sampling? Top-k sampling picks the next token randomly from the k most probable options to add controlled variety to text.
- What Is Top-p Sampling? Top-p sampling selects from the smallest set of tokens whose combined probability exceeds a threshold p.
- What Is Nucleus Sampling? Nucleus sampling is another name for top-p sampling, drawing tokens from a dynamic set of the most probable options.
- What Is Temperature Sampling? Temperature sampling scales token probabilities to control randomness, with higher values producing more varied text.
- What Is a Repetition Penalty? A repetition penalty reduces the probability of already-used tokens to discourage repetitive text generation.
- What Are Logits? Logits are the raw, unnormalized scores a model outputs before they are converted into probabilities.
- What Is a Prompt? A prompt is the input text given to a language model to guide the response it generates.
- What Is a System Prompt? A system prompt sets a model's role, behavior, and constraints before it handles user messages.
- What Is a Prompt Template? A prompt template is a reusable text pattern with placeholders that are filled in to build consistent prompts.
- What Is Zero-Shot Prompting? Zero-shot prompting asks a model to perform a task with only instructions and no worked examples.
- What Is Few-Shot Prompting? Few-shot prompting includes a few examples in the prompt to show a model how to perform a task.
- What Is In-Context Learning? In-context learning lets a model adapt to a task from examples in the prompt without updating its weights.
- What Is Prompt Injection? Prompt injection is an attack where crafted input overrides a model's intended instructions or safeguards.
- What Is Jailbreaking LLMs? Jailbreaking LLMs means crafting prompts that bypass a model's safety guardrails to produce restricted outputs.
- What Is Chain-of-Thought Prompting? Chain-of-thought prompting encourages a model to reason step by step before giving a final answer.
- What Is Tree of Thoughts? Tree of thoughts is a reasoning method where a model explores multiple branching reasoning paths before deciding.
- What Is Self-Consistency? Self-consistency samples multiple reasoning paths and picks the most common answer to improve reliability.
- What Is ReAct Prompting? ReAct prompting interleaves reasoning steps with actions like tool use so a model can solve tasks iteratively.
- What Is Instruction Tuning? Instruction tuning trains a model on many task instructions so it follows natural-language directions better.
- What Is Supervised Fine-Tuning? Supervised fine-tuning adapts a pretrained model using labeled input-output examples for a target behavior.
- What Is Parameter-Efficient Fine-Tuning? Parameter-efficient fine-tuning adapts a large model by training only a small subset of added parameters.
- What Is LoRA? LoRA fine-tunes large models efficiently by learning small low-rank weight updates instead of changing all weights.
- What Is QLoRA? QLoRA combines quantization with LoRA to fine-tune large models on limited memory hardware.
- What Is Prefix Tuning? Prefix tuning adapts a model by learning trainable vectors prepended to each layer while freezing the base weights.
- What Is Prompt Tuning? Prompt tuning learns a set of continuous embeddings that act as a soft prompt while the model stays frozen.
- What Are Adapter Layers? Adapter layers are small trainable modules inserted into a frozen model to adapt it to new tasks efficiently.
- What Is GPTQ? GPTQ is a post-training quantization method that compresses large language models to low-bit precision.
- What Is AWQ? AWQ is an activation-aware quantization method that preserves important weights when compressing language models.
- What Is Model Pruning? Model pruning removes unimportant weights or components from a neural network to make it smaller and faster.
- What Is Flash Attention? Flash attention is an optimized algorithm that computes self-attention faster and with less memory.
- What Is a KV Cache? A KV cache stores past key and value tensors so a model can generate tokens without recomputing prior attention.
- What Is Speculative Decoding? Speculative decoding speeds generation by using a small model to draft tokens that a larger model verifies.
- What Is Context Length? Context length is the maximum number of tokens a language model can process in a single input and output.
- What Is Long Context? Long context refers to models able to process very large inputs, enabling reasoning over lengthy documents.
- What Is Sliding Window Attention? Sliding window attention limits each token to attend to a fixed nearby window, improving efficiency on long inputs.
- What Is Rotary Position Embedding? Rotary position embedding encodes token positions by rotating query and key vectors, aiding generalization to longer sequences.
- What Is ALiBi? ALiBi encodes position by biasing attention scores based on token distance, helping models extend to longer sequences.
- What Is Grouped-Query Attention? Grouped-query attention shares key and value projections across query heads to reduce memory and speed up inference.
- What Is Semantic Search? Semantic search finds results by meaning using vector embeddings rather than exact keyword matching.
- What Is Dense Retrieval? Dense retrieval matches queries to documents using learned embedding vectors and similarity in vector space.
- What Is Sparse Retrieval? Sparse retrieval matches documents using term-based scores like BM25, relying on exact and weighted keyword overlap.
- What Is BM25? BM25 is a ranking function that scores documents by term frequency and rarity for keyword-based search.
- What Is Reranking? Reranking reorders an initial set of retrieved results using a more precise model to improve relevance.
- What Is an Embedding Model? An embedding model converts text into numerical vectors that capture meaning for search and comparison.
- What Is a Sentence Embedding? A sentence embedding is a single vector representing the meaning of an entire sentence or short passage.
- What Is Cosine Similarity? Cosine similarity measures how similar two vectors are by the angle between them, common in text search.
- What Is Named Entity Recognition? Named entity recognition identifies and labels entities like people, places, and organizations in text.
- What Is Part-of-Speech Tagging? Part-of-speech tagging labels each word in a sentence with its grammatical category, such as noun or verb.
- What Is Coreference Resolution? Coreference resolution determines which words or phrases in a text refer to the same entity.
- What Is Dependency Parsing? Dependency parsing analyzes grammatical structure by linking words through head-dependent relationships.
- What Is Text Classification? Text classification assigns predefined categories or labels to a piece of text, such as spam or topic tags.
- What Is Question Answering? Question answering systems return direct answers to questions posed in natural language.
- What Is Text Summarization? Text summarization condenses a document into a shorter version that preserves its key information.
- What Is Machine Translation? Machine translation automatically converts text from one language into another using computational models.
- What Is Language Detection? Language detection identifies which natural language a given piece of text is written in.
- What Are Stop Words? Stop words are common words like 'the' and 'and' often removed before text processing to focus on meaningful terms.
- What Is Lemmatization? Lemmatization reduces words to their dictionary base form using vocabulary and grammar rules.
- What Is Stemming? Stemming trims words to a root form by removing affixes, often using simple rule-based cutting.
- What Is TF-IDF? TF-IDF scores a word's importance in a document by weighing its frequency against how common it is overall.
- What Is Bag of Words? Bag of words represents text by word counts while ignoring word order and grammar.
- What Is an N-gram? An n-gram is a contiguous sequence of n words or characters used in text modeling and analysis.
- What Is a Corpus? A corpus is a large, structured collection of text used to train or evaluate language models.
- What Is Text Generation? Text generation is the production of coherent natural language by a model, often one token at a time.
- What Is a Dialogue System? A dialogue system is software that carries on a conversation with users through text or speech.
- What Is WordPiece Tokenization? WordPiece is a subword tokenization method used by BERT that builds a vocabulary of frequent word pieces.
- What Is the Softmax Function? The softmax function turns a vector of scores into a probability distribution that sums to one.
Vision & Generative Media
- What Is a Convolutional Neural Network? A neural network that uses convolutional layers to detect spatial patterns, widely used for image and visual data tasks.
- What Is Convolution in Deep Learning? A mathematical operation that slides a filter over input data to produce feature maps highlighting local patterns.
- What Is a Pooling Layer? A layer that downsamples feature maps to reduce spatial size and computation while retaining important information.
- What Is Max Pooling? A pooling operation that keeps the maximum value in each region of a feature map to downsample it.
- What Is a Feature Map? The output produced when a filter is applied to input data, representing detected features at each location.
- What Is a Kernel Filter? A small matrix of learnable weights convolved over input data to detect specific visual features.
- What Is Stride in Convolution? The step size a filter moves across input during convolution, controlling output size and overlap.
- What Is Padding in Neural Networks? Adding extra border values around input data so convolution can preserve spatial dimensions.
- What Is Image Classification? The task of assigning a label or category to an entire image based on its content.
- What Is Object Detection? A vision task that locates and labels multiple objects in an image using bounding boxes.
- What Is Image Segmentation? Partitioning an image into regions by assigning a class label to each pixel.
- What Is Semantic Segmentation? Assigning a class label to every pixel without distinguishing separate object instances.
- What Is Instance Segmentation? Detecting objects and labeling each pixel while distinguishing separate instances of the same class.
- What Is a Bounding Box? A rectangular region drawn around an object to indicate its location in an image.
- What Is YOLO Object Detection? A real-time object detection approach that predicts bounding boxes and classes in a single network pass.
- What Is R-CNN? A region-based convolutional network that proposes candidate regions then classifies each for object detection.
- What Is Faster R-CNN? An object detection model that uses a region proposal network for efficient, accurate detection.
- What Is Mask R-CNN? An extension of Faster R-CNN that adds a branch for predicting object segmentation masks.
- What Is an SSD Detector? A single-shot detector that predicts objects at multiple scales in one network pass.
- What Is ResNet? A deep CNN architecture using residual connections to enable training of very deep networks.
- What Is the VGG Network? A deep CNN architecture known for using stacks of small convolutional filters.
- What Is AlexNet? An influential deep CNN that popularized deep learning for image classification.
- What Is the Inception Network? A CNN architecture using inception modules that process input at multiple filter sizes in parallel.
- What Is EfficientNet? A CNN family that scales depth, width, and resolution together for strong efficiency.
- What Is MobileNet? A lightweight CNN architecture designed for efficient use on mobile and edge devices.
- What Is a Vision Transformer? A model that applies transformer architecture to image patches for vision tasks.
- What Is CLIP? A model trained to connect images and text, enabling zero-shot image classification.
- What Is Image Recognition? The ability of a system to identify objects, people, or features within images.
- What Is Facial Recognition? Technology that identifies or verifies a person from their facial features.
- What Is Face Detection? The task of locating human faces within an image or video frame.
- What Is Pose Estimation? Detecting the position and orientation of body joints or objects in images.
- What Is Optical Character Recognition? Technology that converts images of text into machine-readable text.
- What Is Image Captioning? Generating a natural-language description of the content of an image.
- What Is Visual Question Answering? A task where a model answers natural-language questions about an image.
- What Is Edge Detection? An image processing technique that identifies boundaries where brightness changes sharply.
- What Is Image Preprocessing? Preparing raw images through steps like resizing, normalization, and augmentation before analysis.
- What Is Data Labeling? Annotating raw data with labels so machine learning models can learn from it.
- What Is Image Annotation? Marking images with labels, boxes, or masks to create training data for vision models.
- What Is a Generative Adversarial Network? A model with two competing networks, a generator and discriminator, that produce realistic data.
- What Is a Generator Network? The part of a GAN that creates synthetic data to fool the discriminator.
- What Is a Discriminator Network? The part of a GAN that distinguishes real data from generated samples.
- What Is a Conditional GAN? A GAN that generates data conditioned on additional input such as a class label.
- What Is CycleGAN? A GAN that translates images between two domains without paired training examples.
- What Is StyleGAN? A GAN architecture known for generating highly realistic and controllable images, especially faces.
- What Is DCGAN? A GAN that uses convolutional layers for stable image generation.
- What Is a Variational Autoencoder? A generative model that encodes data into a probabilistic latent space and reconstructs it.
- What Is Latent Space? A compressed representation where data is encoded as points capturing key features.
- What Is Latent Diffusion? A diffusion model that operates in a compressed latent space for efficient image generation.
- What Is Denoising Diffusion? A generative process that learns to reverse gradual noise to create data samples.
- What Is Stable Diffusion XL? An open text-to-image diffusion model producing higher-resolution, detailed images.
- What Is DALL-E 3? An OpenAI text-to-image model that generates images from natural-language prompts.
- What Is Imagen? A Google text-to-image diffusion model that generates images from text prompts.
- What Is Midjourney v6? A version of the Midjourney text-to-image service known for stylized, detailed art.
- What Is ControlNet? An add-on that guides diffusion models using structural inputs like edges or poses.
- What Is Image Inpainting? Filling in missing or removed regions of an image in a visually consistent way.
- What Is Image Outpainting? Extending an image beyond its original borders with newly generated content.
- What Is Super-Resolution? Increasing the resolution and detail of an image beyond its original quality.
- What Is Image Upscaling? Enlarging an image while preserving or enhancing detail using AI methods.
- What Is Style Transfer? Applying the visual style of one image to the content of another.
- What Is Neural Style Transfer? A deep learning method that recombines the content of one image with the style of another.
- What Is Image-to-Image Generation? Transforming an input image into a new image guided by a model or prompt.
- What Is Text-to-3D Generation? Creating three-dimensional models or scenes from natural-language descriptions.
- What Is a NeRF? A neural method that represents 3D scenes for novel-view rendering from images.
- What Is Gaussian Splatting? A technique that represents 3D scenes with many small blobs for fast rendering.
- What Is 3D Generation? Creating three-dimensional assets such as models or scenes using AI techniques.
- What Is Video Generation? Producing video clips using AI models, from prompts, images, or other inputs.
- What Is Text-to-Video Generation? Creating video clips directly from natural-language text prompts.
- What Is Frame Interpolation? Generating intermediate frames between existing ones to smooth or slow motion.
- What Is Motion Capture? Recording the movement of people or objects to animate digital models.
- What Is Deepfake Detection? Techniques for identifying synthetic or manipulated media created by AI.
- What Is Face Swapping? Replacing a person's face in an image or video with another face.
- What Is AI Image Editing? Using AI tools to modify images through prompts, masks, or automated adjustments.
- What Is Background Removal? Automatically separating a subject from its background in an image.
- What Is Generative Fill? An AI editing feature that fills selected areas with contextually generated content.
- What Is Prompt-to-Image Generation? Producing images directly from descriptive text prompts using generative models.
- What Is a Negative Prompt? A prompt specifying what to avoid or exclude in a generated image.
- What Is a Seed in Image Generation? A number that initializes randomness so image generation can be reproduced.
- What Is Guidance Scale? A setting controlling how closely a generated image follows its text prompt.
- What Are Diffusion Steps? The number of denoising iterations used to generate an image with a diffusion model.
- What Is a Sampler in Diffusion Models? An algorithm that controls how a diffusion model steps through denoising to form an image.
- What Is a VAE Decoder? The component that converts latent representations back into pixel images.
- What Is Text-to-Speech? Technology that converts written text into spoken audio.
- What Is Speech Synthesis? The artificial production of human-like speech from text or other inputs.
- What Is Voice Cloning Technology? Creating a synthetic copy of a specific person's voice from samples.
- What Is Voice Conversion? Transforming one speaker's voice to sound like another while keeping the words.
- What Is Audio Generation? Creating audio such as speech, music, or sound effects using AI models.
- What Is Music Generation? Composing or producing music using AI models from prompts or examples.
- What Is Sound Effect Generation? Producing sound effects with AI from text or other input descriptions.
- What Is Automatic Speech Recognition? Technology that converts spoken language into written text.
- What Is Speech-to-Text? Converting spoken words into written text automatically.
- What Is the Whisper Model? An OpenAI speech recognition model that transcribes and translates audio.
- What Is Speaker Diarization? Determining who spoke when by segmenting audio by speaker.
- What Is Audio Classification? Assigning category labels to audio clips based on their content.
- What Is a Mel Spectrogram? A visual representation of audio frequencies scaled to human hearing perception.
- What Is WaveNet? A deep neural network that generates raw audio waveforms for realistic speech.
- What Is a Vocoder? A component that converts acoustic features into an audible speech waveform.
- What Is Lip-Sync AI? Technology that aligns a speaker's mouth movements with given audio.
- What Is Avatar Generation? Creating digital characters or avatars, sometimes animated, using AI.
- What Is a Talking Head Model? AI that animates a still portrait to speak in sync with audio.
- What Is Video Upscaling? Increasing the resolution and detail of video using AI methods.
AI in Industry
- How Is AI Used in Finance? AI in finance applies machine learning to trading, risk, fraud detection, and customer service across banking and investment firms.
- What Is Algorithmic Trading? Algorithmic trading uses computer programs to execute financial trades automatically based on predefined rules and market data.
- How Does AI Detect Fraud? AI fraud detection uses machine learning to spot unusual patterns in transactions that may indicate fraudulent activity.
- What Is AI Credit Scoring? AI credit scoring uses machine learning to assess a borrower's creditworthiness from financial and behavioral data.
- What Is a Robo-Advisor? A robo-advisor is an automated platform that provides algorithm-driven investment advice and portfolio management with little human input.
- How Is AI Used in Banking? AI in banking supports fraud detection, customer service chatbots, credit decisions, and process automation across financial services.
- How Is AI Used in Insurance? Insurance AI applies machine learning to underwriting, risk assessment, claims processing, and fraud detection.
- How Is AI Used in Marketing? AI in marketing analyzes customer data to personalize campaigns, target audiences, and optimize advertising spend.
- What Is Personalized Marketing? Personalized marketing uses AI to tailor content, offers, and recommendations to individual customers based on their data.
- How Is AI Used in Advertising? AI advertising automates ad targeting, creative optimization, and budget allocation using machine learning and data analysis.
- What Is AI Customer Segmentation? AI customer segmentation groups customers by shared traits using machine learning to enable targeted marketing.
- How Is AI Used in Retail? AI in retail supports demand forecasting, inventory management, personalized recommendations, and automated checkout.
- What Is AI Demand Forecasting? AI demand forecasting predicts future product demand using historical sales data and machine learning models.
- What Is AI Inventory Optimization? AI inventory optimization uses forecasting and analytics to maintain the right stock levels and reduce holding costs.
- What Is Dynamic Pricing? Dynamic pricing uses AI to adjust prices in real time based on demand, competition, and other market factors.
- How Is AI Used in E-Commerce? AI in e-commerce powers product recommendations, search, personalized pricing, and customer support for online stores.
- What Is a Product Recommendation System? A product recommendation system suggests items to shoppers using AI based on their behavior and preferences.
- How Is AI Used in Manufacturing? AI in manufacturing supports predictive maintenance, quality inspection, and automation to improve production efficiency.
- What Is Predictive Maintenance? Predictive maintenance uses AI to anticipate equipment failures before they happen, reducing downtime and repair costs.
- How Is AI Used for Quality Inspection? Quality inspection AI uses computer vision to detect product defects on production lines automatically.
- What Is Industrial Automation? Industrial automation uses control systems and AI to operate machinery and processes with minimal human intervention.
- How Is AI Used in Supply Chain? AI in supply chain improves forecasting, logistics, and inventory decisions through data analysis and automation.
- What Is AI Logistics Optimization? AI logistics optimization improves the movement and storage of goods using data-driven planning and routing.
- What Is Route Optimization? Route optimization uses AI to find the most efficient delivery or travel routes based on multiple constraints.
- What Is Warehouse Robotics? Warehouse robotics uses AI-guided robots to move, sort, and store goods in fulfillment and distribution centers.
- How Is AI Used in Agriculture? AI in agriculture supports crop monitoring, precision farming, and yield prediction using sensors and imagery.
- What Is Precision Agriculture? Precision agriculture uses AI, sensors, and data to manage crops and resources at a fine, field-specific level.
- What Is AI Crop Monitoring? AI crop monitoring uses imagery and sensors to track crop health, growth, and stress across fields.
- How Is AI Used in Medical Imaging? Medical imaging AI helps analyze X-rays, MRIs, and scans to support diagnosis and detect abnormalities.
- How Is AI Used in Drug Discovery? AI in drug discovery accelerates finding and designing new medicines by analyzing molecular and biological data.
- What Is AI Protein Folding? AI protein folding predicts the 3D structure of proteins from their amino acid sequences to aid biology and medicine.
- What Is AlphaFold? AlphaFold is an AI system developed by DeepMind that predicts protein 3D structures from amino acid sequences.
- How Is AI Used in Radiology? Radiology AI analyzes medical images to help detect conditions and support radiologists in diagnosis.
- What Is Clinical Decision Support? Clinical decision support uses AI to provide clinicians with data-driven guidance during patient care.
- How Is AI Used in Mental Health? AI in mental health supports screening, chatbots, and monitoring tools to help expand access to care.
- How Is AI Used in Pharma? AI in pharma supports drug discovery, clinical trial design, and manufacturing across the pharmaceutical industry.
- What Is an Intelligent Tutoring System? An intelligent tutoring system uses AI to give students personalized instruction, feedback, and practice.
- What Is Automated Grading? Automated grading uses AI to evaluate student work such as essays and assignments, providing scores and feedback.
- How Is AI Used in Law? AI in law supports legal research, document review, and contract analysis for lawyers and legal teams.
- What Is Legal Document Analysis? Legal document analysis uses AI to review, search, and summarize legal texts to speed up legal work.
- What Is Contract Analysis AI? Contract analysis AI reviews contracts to identify clauses, risks, and obligations using natural language processing.
- How Is AI Used in HR? AI in HR supports recruiting, resume screening, employee analytics, and workforce planning.
- What Is AI Resume Screening? AI resume screening automatically reviews job applications to identify candidates matching role requirements.
- How Is AI Used in Recruiting? AI recruiting uses automation and data analysis to source, screen, and engage job candidates.
- How Is AI Used in Customer Service? AI in customer service powers chatbots, virtual assistants, and automation to handle support requests.
- What Is a Virtual Assistant? A virtual assistant is an AI program that responds to voice or text requests to perform tasks and answer questions.
- What Is Conversational AI? Conversational AI enables machines to understand and respond to human language in natural, dialogue-based interactions.
- How Is AI Used in Call Centers? Call center AI automates calls, routes inquiries, and assists agents using speech recognition and NLP.
- How Is AI Used in Cybersecurity? AI in cybersecurity detects threats, analyzes network activity, and automates responses to security incidents.
- What Is AI Threat Detection? AI threat detection identifies potential security threats by analyzing patterns in data and system activity.
- What Is AI Intrusion Detection? AI intrusion detection monitors networks and systems to identify unauthorized access and suspicious behavior.
- How Is AI Used in Malware Detection? AI malware detection identifies malicious software by analyzing code behavior and patterns with machine learning.
- How Does AI Filter Spam? AI spam filtering uses machine learning to identify and block unwanted or malicious email messages.
- How Is AI Used in Gaming? AI in gaming controls non-player characters, generates content, and adapts gameplay to players.
- What Is Game AI? Game AI refers to techniques that make characters and systems in video games behave intelligently.
- What Is Procedural Generation? Procedural generation uses algorithms to automatically create game content such as levels, maps, and worlds.
- What Is NPC Behavior in AI? NPC behavior refers to how AI controls non-player characters to act and react within a video game.
- How Is AI Used in Transportation? AI in transportation supports traffic prediction, route planning, autonomous vehicles, and driver assistance.
- What Is AI Traffic Prediction? AI traffic prediction forecasts road conditions and congestion using historical and real-time data.
- What Are Autonomous Vehicles? Autonomous vehicles use AI, sensors, and cameras to navigate and drive with reduced or no human input.
- What Is ADAS? ADAS, or advanced driver assistance systems, use AI and sensors to help drivers with safety and control tasks.
- How Is AI Used in Energy? AI in energy supports demand forecasting, grid management, and efficiency across power generation and distribution.
- What Is a Smart Grid? A smart grid uses digital technology and AI to monitor and manage electricity supply and demand efficiently.
- What Is AI Energy Forecasting? AI energy forecasting predicts electricity demand and renewable generation to help balance the grid.
- How Is AI Used in Real Estate? AI in real estate supports property valuation, market analysis, and personalized property search.
- What Is Property Valuation AI? Property valuation AI estimates real estate prices using data on location, features, and comparable sales.
- How Is AI Used in Journalism? AI in journalism supports automated reporting, research, and content personalization for news organizations.
- What Is Automated Journalism? Automated journalism uses AI to generate news articles from structured data, such as sports or financial reports.
- How Is AI Used in the Music Industry? AI in the music industry supports composition, production, recommendation, and audio analysis.
- How Is AI Used in Film? AI in film supports visual effects, editing, dubbing, and content generation for the movie industry.
- How Is AI Used in Fashion? AI in fashion supports trend forecasting, design, personalized recommendations, and virtual try-on.
- How Is AI Used in Sports? AI in sports supports performance analysis, injury prevention, scouting, and fan engagement.
- What Is Sports Analytics? Sports analytics uses data and AI to evaluate performance, inform strategy, and support team decisions.
- How Is AI Used in Government? AI in government supports public services, fraud detection, policy analysis, and smart city initiatives.
- What Is a Smart City? A smart city uses AI, sensors, and data to manage services like traffic, energy, and public safety efficiently.
- How Is AI Used in Climate Science? AI in climate science supports modeling, emissions tracking, and analysis of environmental data.
- What Is AI Climate Modeling? AI climate modeling uses machine learning to simulate and predict climate patterns and environmental change.
- How Is AI Used in Weather Forecasting? Weather forecasting AI uses machine learning to predict weather conditions from atmospheric and sensor data.
- How Is AI Used in Space? AI in space supports spacecraft autonomy, satellite operations, and analysis of astronomical data.
- What Is Satellite Imagery Analysis? Satellite imagery analysis uses AI to interpret images from space for mapping, monitoring, and detection.
- What Is AI Remote Sensing? AI remote sensing analyzes data from satellites and sensors to monitor Earth's surface and environment.
- What Are Collaborative Robots (Cobots)? Collaborative robots, or cobots, work safely alongside humans using sensors and AI in shared workspaces.
- How Is AI Used in Telecommunications? AI in telecommunications supports network optimization, fault detection, and customer service automation.
- What Is AI Network Optimization? AI network optimization improves the performance and reliability of communication networks using data analysis.
- What Is Churn Prediction? Churn prediction uses AI to identify customers likely to stop using a product or service so businesses can retain them.
- What Is AI Lead Scoring? AI lead scoring ranks sales prospects by their likelihood to convert using data and machine learning.
- How Is AI Used in Accounting? AI in accounting automates data entry, reconciliation, expense processing, and anomaly detection.
- What Is Intelligent Document Processing? Intelligent document processing uses AI to extract and organize data from documents like invoices and forms.
- How Is AI Used in Construction? AI in construction supports project planning, safety monitoring, design, and progress tracking.
- How Is AI Used in Mining? AI in mining supports exploration, equipment automation, safety monitoring, and predictive maintenance.
- How Is AI Used in Oil and Gas? AI in oil and gas supports exploration, drilling optimization, equipment monitoring, and safety.
- How Is AI Used in the Food Industry? AI in the food industry supports quality control, safety inspection, demand forecasting, and production.
- What Is a Personalization Engine? A personalization engine uses AI to tailor content, products, and experiences to individual users.
- How Is AI Used in Translation? AI translation services use machine learning to convert text and speech between languages automatically.
- How Is AI Used in Insurance Claims? AI in insurance claims automates processing, assesses damage, and detects fraud to speed up settlements.
- How Is AI Used in Tax? AI in tax automates data processing, compliance checks, and analysis to streamline tax preparation.
- How Is AI Used in Logistics? AI in logistics optimizes routing, warehousing, and delivery to move goods efficiently across supply chains.
- How Is AI Used in Retail Checkout? AI in retail checkout enables cashierless stores and automated payment using computer vision and sensors.
- What Is AI in Advertising Bidding? AI in advertising bidding automates real-time bids for online ad placements to maximize campaign value.
- How Is Computer Vision Used in Industry? Industrial computer vision uses AI to analyze images for inspection, automation, and monitoring across many sectors.
Infrastructure & Agents
- What Is a Graphics Processing Unit (GPU)? A GPU is a processor with many parallel cores, widely used to accelerate the matrix math behind neural network training and inference.
- What Is a Tensor Processing Unit (TPU)? A TPU is a Google-designed chip built specifically to accelerate the tensor operations used in machine learning workloads.
- What Is CUDA? CUDA is NVIDIA's parallel computing platform and API that lets developers run general-purpose code on GPUs.
- What Is a GPU Cluster? A GPU cluster is a group of networked machines with many GPUs working together to train or serve large AI models.
- What Is an AI Accelerator? An AI accelerator is specialized hardware designed to speed up machine learning computation compared to general-purpose CPUs.
- What Is a Neural Processing Unit (NPU)? An NPU is a processor specialized for neural network operations, often built into phones and laptops for efficient on-device AI.
- What Is Distributed Training? Distributed training splits a machine learning workload across multiple devices or machines to train models faster or at larger scale.
- What Is Data Parallelism? Data parallelism trains copies of a model on different data batches across devices, then combines the resulting gradients.
- What Is Model Parallelism? Model parallelism splits a single model's layers or parameters across multiple devices when it is too large to fit on one.
- What Is Tensor Parallelism? Tensor parallelism splits individual tensor operations, such as large matrix multiplications, across multiple devices.
- What Is Pipeline Parallelism? Pipeline parallelism assigns different model layers to different devices and streams data through them like an assembly line.
- What Is Mixed-Precision Training? Mixed-precision training uses lower-precision numbers for most computation while keeping higher precision where accuracy matters.
- What Is FP16 (Half Precision)? FP16 is a 16-bit floating-point format used to reduce memory use and speed up neural network computation.
- What Is bfloat16? bfloat16 is a 16-bit floating-point format with a wide numeric range, commonly used in machine learning hardware.
- What Is INT8 Quantization? INT8 quantization represents model weights and activations as 8-bit integers to shrink models and speed up inference.
- What Is Model Serving? Model serving is the practice of hosting a trained model so applications can send inputs and receive predictions in production.
- What Is Inference Optimization? Inference optimization improves the speed, cost, or memory use of running trained models to generate predictions.
- What Is Batch Inference? Batch inference processes many inputs together in groups, improving throughput when immediate results are not required.
- What Is Real-Time Inference? Real-time inference generates predictions with low latency so results are returned almost immediately to users or systems.
- What Is Model Deployment? Model deployment is the process of moving a trained model into a production environment where it can serve real requests.
- What Is Edge AI? Edge AI runs machine learning models on local devices near the data source rather than in a centralized cloud.
- What Is Edge Computing for AI? Edge computing for AI processes model workloads close to where data is generated to reduce latency and reliance on the cloud.
- What Is On-Device AI? On-device AI runs models directly on a user's device, keeping data local and enabling offline functionality.
- What Is TinyML? TinyML brings machine learning to very small, low-power devices such as microcontrollers and embedded sensors.
- What Is Federated Learning? Federated learning trains a shared model across many devices without moving their raw data to a central server.
- What Is MLOps? MLOps is a set of practices for reliably building, deploying, and maintaining machine learning systems in production.
- What Is Model Monitoring? Model monitoring tracks a deployed model's performance and behavior to detect issues like accuracy loss over time.
- What Is Model Drift? Model drift is the gradual decline in a model's accuracy as real-world conditions diverge from its training data.
- What Is Data Drift? Data drift occurs when the statistical properties of incoming data change compared to the data a model was trained on.
- What Is a Model Registry? A model registry is a central catalog for storing, versioning, and managing trained machine learning models.
- What Is a Feature Store? A feature store is a system for storing, sharing, and serving the input features used by machine learning models.
- What Is Experiment Tracking? Experiment tracking records the settings, code, data, and results of machine learning runs so they can be compared and reproduced.
- What Is Model Versioning? Model versioning tracks different iterations of a model so teams can compare, roll back, and reproduce them.
- What Is CI/CD for Machine Learning? CI/CD for ML automates testing, building, and deploying machine learning models and pipelines to speed up delivery.
- What Is Kubernetes for Machine Learning? Kubernetes is a container orchestration platform often used to deploy and scale machine learning workloads.
- What Is Docker for Machine Learning? Docker packages machine learning code and dependencies into containers so they run consistently across environments.
- What Is the Ray Framework? Ray is an open-source framework for scaling Python and machine learning workloads across multiple machines.
- What Is a Model Checkpoint? A model checkpoint is a saved snapshot of a model's parameters during training, used to resume or restore it later.
- What Is an ML Pipeline? An ML pipeline is a sequence of automated steps that move data through preparation, training, and deployment.
- What Is a Data Pipeline? A data pipeline is a series of steps that move and transform data from sources to destinations for analysis or model training.
- What Is ETL (Extract, Transform, Load)? ETL is a process that extracts data from sources, transforms it into a usable form, and loads it into a target system.
- What Is a Data Lake? A data lake is a storage system that holds large amounts of raw data in many formats until it is needed.
- What Is a Data Warehouse? A data warehouse is a system that stores structured, organized data optimized for analysis and reporting.
- What Is Vector Search? Vector search finds items whose numeric embeddings are closest to a query, enabling semantic similarity retrieval.
- What Is Pinecone? Pinecone is a managed vector database service used to store embeddings and perform fast similarity search.
- What Is Weaviate? Weaviate is an open-source vector database for storing embeddings and running semantic search over data.
- What Is FAISS? FAISS is an open-source library from Meta for efficient similarity search over large collections of vectors.
- What Is Chroma DB? Chroma is an open-source embedding database designed to make it easy to store and query vectors for AI apps.
- What Is Milvus? Milvus is an open-source vector database built for storing and searching large-scale embedding data.
- What Is an AI Agent Framework? An AI agent framework provides tools and structure for building software agents that plan, use tools, and act autonomously.
- What Is LangChain? LangChain is an open-source framework for building applications that combine language models with tools, data, and workflows.
- What Is LlamaIndex? LlamaIndex is a framework for connecting language models to external data through indexing and retrieval.
- What Is AutoGPT? AutoGPT is an open-source project that chains language model calls to pursue goals with limited human input.
- What Is BabyAGI? BabyAGI is an open-source example of a task-driven autonomous agent that generates and prioritizes its own tasks.
- What Is CrewAI? CrewAI is a framework for building systems of multiple AI agents that collaborate on tasks with defined roles.
- What Is an Autonomous Agent? An autonomous agent is an AI system that plans and takes actions toward goals with little ongoing human direction.
- What Is a Multi-Agent System? A multi-agent system uses several AI agents that interact or cooperate to solve problems that are hard for one agent alone.
- What Is Agent Orchestration? Agent orchestration coordinates how multiple AI agents or steps work together to complete a larger task.
- What Is Tool Use in AI Agents? Tool use lets AI models call external functions, APIs, or services to access data and perform actions beyond text generation.
- What Is Function Calling? Function calling lets a language model output structured requests to run predefined functions with specific arguments.
- What Is the Model Context Protocol (MCP)? The Model Context Protocol is an open standard for connecting AI models to external tools and data sources.
- What Is an Agentic Workflow? An agentic workflow structures an AI task as a loop of planning, acting, and observing to reach a goal.
- What Is AI Planning? AI planning is the process by which an agent chooses a sequence of actions to achieve a specified goal.
- What Is Task Decomposition? Task decomposition breaks a complex goal into smaller subtasks that an AI agent can tackle step by step.
- What Is Agent Memory? Agent memory lets an AI agent store and recall information across steps or sessions to maintain context.
- What Is a RAG Pipeline? A RAG pipeline retrieves relevant documents and feeds them to a language model to ground its responses in external data.
- What Is a Semantic Cache? A semantic cache stores past query results and reuses them when a new query is similar in meaning.
- What Is Prompt Caching? Prompt caching reuses previously processed prompt content to lower cost and latency for repeated requests.
- What Are AI Guardrails? Guardrails are controls that constrain AI model inputs and outputs to keep behavior safe, on-topic, and policy-compliant.
- What Is LLM Evaluation? LLM evaluation measures the quality, accuracy, and safety of large language model outputs using tests and metrics.
- What Is LLM Observability? LLM observability provides visibility into how language model applications behave in production, including inputs and outputs.
- What Is Hallucination Detection? Hallucination detection identifies when a language model produces false or unsupported information in its output.
- What Is an AI API? An AI API is an interface that lets developers send requests to AI models and receive predictions or generated content.
- What Is the OpenAI API? The OpenAI API lets developers access OpenAI's models for tasks like text generation, embeddings, and image creation.
- What Is the Anthropic API? The Anthropic API provides developers access to Anthropic's Claude models for text and reasoning tasks.
- What Is API Rate Limiting? API rate limiting restricts how many requests a client can make in a given period to protect service stability.
- What Is Token Cost? Token cost is the price charged per unit of text, measured in tokens, when using many language model APIs.
- What Is Inference Cost? Inference cost is the expense of running a trained model to generate predictions, driven by compute and usage.
- What Is Model Hosting? Model hosting is the practice of running a model on infrastructure so applications can access it over a network.
- What Is Serverless AI? Serverless AI runs models without managing servers, scaling automatically and charging based on usage.
- What Is AWS SageMaker? AWS SageMaker is Amazon's cloud service for building, training, and deploying machine learning models.
- What Is Azure AI? Azure AI is Microsoft's set of cloud services and tools for building and running AI and machine learning solutions.
- What Is Google Vertex AI? Google Vertex AI is Google Cloud's unified platform for building, training, and deploying machine learning models.
- What Is Hugging Face Transformers? Transformers is a popular open-source library from Hugging Face for using and fine-tuning pretrained models.
- What Is PyTorch? PyTorch is an open-source deep learning framework widely used for research and production machine learning.
- What Is TensorFlow? TensorFlow is an open-source machine learning framework developed by Google for building and deploying models.
- What Is Keras? Keras is a high-level neural network API that makes building and training deep learning models simpler.
- What Is JAX? JAX is a Google library for high-performance numerical computing and machine learning with automatic differentiation.
- What Is ONNX? ONNX is an open format for representing machine learning models so they can be shared across frameworks and tools.
- What Is GGUF? GGUF is a file format for storing language models, commonly used for running quantized models locally.
- What Is Ollama? Ollama is a tool for running open large language models locally on your own computer with a simple interface.
- What Is llama.cpp? llama.cpp is an open-source project for running language models efficiently on CPUs and consumer hardware.
- What Is vLLM? vLLM is an open-source library for serving large language models with high throughput and efficient memory use.
- What Is a Retrieval Pipeline? A retrieval pipeline finds and ranks relevant documents from a data source to supply context to an AI model.
- What Is an Embedding Pipeline? An embedding pipeline converts data into numeric vectors and stores them for similarity search and retrieval.
- What Are Model Quantization Tools? Model quantization tools reduce a model's numeric precision to shrink its size and speed up inference.
- What Is Gradient Checkpointing? Gradient checkpointing saves memory during training by recomputing some intermediate values instead of storing them.
- What Is Activation Checkpointing? Activation checkpointing reduces training memory by storing fewer activations and recomputing them when needed.
- What Is the Zero Redundancy Optimizer (ZeRO)? ZeRO is a technique that partitions optimizer state and model data across devices to train very large models efficiently.
- What Is DeepSpeed? DeepSpeed is a Microsoft open-source library that optimizes large-scale deep learning training and inference.