Skip to main content
Zubnet AI › Learn › Wiki

AI Wiki

AI concepts explained by builders, not textbooks. No jargon walls. No academic gatekeeping. Just clear, practical definitions of the terms you'll actually encounter.

375 terms 8 categories Updated 2026-10-14
💡
Term of the Day
Automation
Using AI to perform tasks that previously required human intervention. This ranges from simple automation (auto-categorizing emails, generating…
→
🧭 Learning Paths ▼
Beginner
I just heard about AI
AI→Chatbot→Prompt→LLM→Token→Context Window→Hallucination
Builder
I'm building an AI app
API→Structured Output→Streaming→Function Calling→RAG→Semantic Search→Model Serving
Deep Dive
How does AI actually work?
Neuron→Layer→Activation Function→Gradient Descent→Transformer→Attention→Autoregressive
Local AI
I want to run AI on my machine
Open Weights→Quantization→GGUF→llama.cpp→Ollama→VRAM→Edge AI
Safety
AI safety & alignment
Alignment→Guardrails→Red Teaming→Prompt Injection→Constitutional AI→AI Ethics→AI Regulation
ML Engineer
I want to train models
Dataset→Loss Function→Transfer Learning→Fine-Tuning→LoRA→RLHF→DPO
No terms match your search.
01.AI01.AI, Yi
Companies
A Chinese AI company founded by Kai-Fu Lee in 2023, best known for its Yi series of large language models. 01.AI…
A/B Testing for AIOnline Evaluation, Split Testing
infra
Comparing two AI system variants (different models, prompts, or configurations) by randomly assigning real users to…
A2AAgent2Agent, Agent2Agent Protocol
Tools
A2A (Agent2Agent) is an open protocol, launched by Google in 2025, that lets AI agents built by different vendors and…
Activation FunctionReLU, GELU, SiLU, Swish
fundamentals
A mathematical function applied to a neuron's output that introduces non-linearity into the network. Without activation…
Adam OptimizerAdam, AdamW
training
The most widely used optimization algorithm for training neural networks. Adam (Adaptive Moment Estimation) combines…
AgentAI Agent
Tools
An AI system that can autonomously plan and execute multi-step tasks, using tools (web search, code execution, API…
Agentic WorkflowAgent Architecture, AI Workflow
using
A design pattern where AI agents orchestrate multi-step processes — planning, executing tools, evaluating…
AGIArtificial General Intelligence
Fundamentals
A hypothetical AI system that can understand, learn, and perform any intellectual task that a human can — with…
AI BenchmarksMMLU, HumanEval, ARC, HellaSwag
fundamentals
Standardized tests used to measure and compare AI model capabilities. MMLU tests knowledge across 57 academic subjects.…
AI Coding AssistantsCode Copilot, AI IDE
Tools
AI tools that help developers write, review, debug, and deploy code. From autocomplete (GitHub Copilot, Codeium) to…
AI Energy ConsumptionAI Power Consumption, AI Energy Use
Infrastructure
The total electricity consumed by AI systems across their lifecycle, dominated by two phases: the one-time training run…
AI EthicsResponsible AI, Ethical AI
safety
The study of moral questions raised by AI development and deployment: What biases do AI systems perpetuate? Who is…
AI GovernanceAI Regulation, AI Policy
Safety
The frameworks, policies, laws, and organizational practices that guide how AI is developed, deployed, and used. This…
AI in CybersecurityCybersecurity AI, AI Threat Detection
Safety
The dual application of AI in cybersecurity: using AI to defend systems (threat detection, anomaly detection, automated…
AI InfrastructureAI Infra, ML Infrastructure
Infrastructure
The full stack of hardware, software, and services required to train and deploy AI models at scale. This includes GPUs…
AI ObservabilityLLM Monitoring, AI Tracing, LLMOps
infra
Monitoring and understanding the behavior of AI systems in production — tracking inputs, outputs, latency, costs…
AI PricingToken Pricing, API Pricing
Infrastructure
How AI providers charge for access to their models. The dominant model is per-token pricing — you pay for the…
AI PrivacyData Privacy in AI, ML Privacy
Safety
The challenge of building and using AI systems without compromising personal data. This spans the entire lifecycle…
AI RegulationEU AI Act, AI Policy
safety
Laws and policies governing the development and deployment of AI systems. The EU AI Act (2024) is the most…
AI Safety
Safety
The field of research and engineering aimed at making AI systems behave reliably, predictably, and in line with human…
AI SecurityLLM Security, AI Safety Engineering
Safety
The practice of protecting AI systems from adversarial attacks, data poisoning, prompt injection, model theft, and…
AI Winter
fundamentals
A period of reduced funding, interest, and progress in AI research following a cycle of hype and unmet expectations.…
AI21 LabsJamba
companies
An Israeli AI company known for Jamba, the first production-grade hybrid architecture that combines Transformer…
Aleph AlphaAleph Alpha
Companies
A German AI company founded in Heidelberg in 2019, best known for its Luminous family of large language models and its…
AlexNet
models
The convolutional neural network that won the 2012 ImageNet competition by a massive margin, triggering the deep…
Alibaba CloudQwen models, Tongyi Qianwen
Companies
The cloud computing arm of Alibaba Group and creator of the Qwen model family. Qwen models are fully open-weights…
Alignment
Safety
The challenge of making AI systems behave in ways that match human values and intentions. An aligned model does what…
AlphaFoldAlphaFold 2, AlphaFold 3
Models
A protein structure prediction system from Google DeepMind that computes the three-dimensional shape of a protein…
AlphaGoAlphaZero, AlphaGo Zero
Models
A Go-playing AI system built by Google DeepMind that in March 2016 became the first computer program to defeat a top…
AnnotationData Labeling, Data Annotation
training
The process of adding labels, tags, or metadata to raw data so it can be used for supervised learning. Annotating…
AnthropicClaude, Constitutional AI, MCP
Companies
AI safety company building Claude. Founded by former OpenAI researchers Dario and Daniela Amodei, Anthropic focuses on…
APIApplication Programming Interface
Infrastructure
A structured way for software to talk to other software. In AI, this usually means sending a request (your prompt) to a…
Apple Intelligence
companies
Apple's on-device and cloud AI system, integrated across iPhone, iPad, and Mac. Apple Intelligence runs smaller models…
Artificial IntelligenceAI, Machine Intelligence
fundamentals
The broad field of building machines that can perform tasks typically requiring human intelligence &mdash…
ASIArtificial Superintelligence
Fundamentals
A theoretical AI system that surpasses the cognitive abilities of all humans in virtually every domain &mdash…
AssemblyAIUniversal-2 STT, audio intelligence
Companies
Speech AI company building developer-friendly APIs for transcription, speaker detection, and audio understanding. Their…
AttentionAttention Mechanism, Self-Attention
Models
The core mechanism in Transformers that lets a model weigh which parts of the input are most relevant to each other.…
Attention VisualizationAttention Maps, Attention Heatmap
fundamentals
Visualizing what a Transformer model "attends to" by displaying the attention weights as heatmaps. For each query…
AutomationAI Automation, Workflow Automation
Tools
Using AI to perform tasks that previously required human intervention. This ranges from simple automation…
Autonomous AgentAI Agent, Agentic AI
using
An AI system that can independently plan, decide, and execute multi-step tasks with minimal human supervision. Given a…
AutoregressiveNext-Token Prediction
fundamentals
A model that generates output one token at a time, where each new token is predicted based on all the tokens that came…
AWS BedrockAmazon Bedrock
companies
Amazon Web Services' managed platform for accessing and deploying foundation models from multiple providers (Anthropic…
BackpropagationBackprop, Backward Pass
fundamentals
The algorithm that computes how much each parameter in a neural network contributed to the error, enabling gradient…
BaiduErnie, Wenxin Yiyan, Ernie Bot
Companies
Chinese technology giant best known for its search engine and for Ernie (Wenxin Yiyan), the chatbot and…
Batch Size & EpochMini-Batch, Training Epoch
training
Batch size is how many training examples the model processes before updating its parameters. An epoch is one complete…
Benchmark
Training
A standardized test used to evaluate and compare AI models. Benchmarks measure specific capabilities — reasoning…
BERTBidirectional Encoder Representations from Transformers
models
A Transformer-based model from Google (2018) that revolutionized NLP by introducing bidirectional pre-training &mdash…
Bias
Safety
Systematic patterns in AI outputs that reflect or amplify societal prejudices present in training data. Bias can appear…
Black Forest LabsFLUX.1 models
Companies
Founded by the original creators of Stable Diffusion after leaving Stability AI. Their FLUX models quickly became the…
BLEU & ROUGEBLEU Score, ROUGE Score
fundamentals
Classic metrics for evaluating text generation quality by comparing model output to reference texts. BLEU (Bilingual…
BPEByte Pair Encoding, Subword Tokenization
fundamentals
The most common algorithm for building tokenizer vocabularies. BPE starts with individual bytes or characters and…
BriaLicensed training data, enterprise image generation
Companies
Israeli AI company that built its image generation models exclusively on licensed, attributed training data. Positions…
ByteDanceDoubao, TikTok, AI-powered recommendation
Companies
Parent company of TikTok and one of the world's most valuable tech companies. Their AI lab builds the Doubao model…
CartesiaSonic, SSM-based voice models
Companies
Voice AI startup built on state space model (SSM) architecture rather than transformers. Their Sonic models achieve…
Catastrophic ForgettingCatastrophic Interference
training
When a neural network trained on a new task loses its ability to perform previously learned tasks. Fine-tuning a model…
CerebrasCerebras Systems, WSE
companies
A chip company that builds wafer-scale AI processors — chips the size of an entire silicon wafer, over 100x…
Chain of ThoughtCoT
Using AI
A prompting technique where you ask the model to show its reasoning step by step before giving a final answer. Instead…
Character.AIc.ai
companies
A platform for creating and chatting with AI characters — fictional personalities, historical figures, and custom…
ChatbotAI Assistant
using
A software interface that lets you interact with an AI model through conversation. Modern AI chatbots (Claude, ChatGPT…
Chatbot ArenaLMSYS Arena, ELO Rankings
fundamentals
A crowdsourced platform (by LMSYS) where users chat with two anonymous AI models side-by-side and vote for which…
CheckpointModel Checkpoint, Snapshot
training
A saved snapshot of a model's state during training — the weights, optimizer state, learning rate schedule, and…
ClassificationClassifier, Categorization
fundamentals
The task of assigning an input to one of a predefined set of categories. "Is this email spam or not?" (binary…
ClaudeClaude Sonnet, Claude Opus, Claude Haiku
Models
Claude is a family of large language models built by Anthropic, offered in three tiers — Haiku, Sonnet, and Opus…
CLIPContrastive Language-Image Pre-training
models
A model from OpenAI (2021) that learns to connect images and text by training on 400 million image-caption pairs. CLIP…
ClusteringK-Means, DBSCAN, Cluster Analysis
fundamentals
An unsupervised learning task that groups similar data points together without predefined labels. Given customer…
CNNConvolutional Neural Network, ConvNet
models
A neural network architecture designed to process grid-like data (images, audio spectrograms) by sliding small filters…
CohereCommand, Embed, Rerank
Companies
Enterprise-focused AI company co-founded by Aidan Gomez, one of the co-authors of the original "Attention Is All You…
Computer UseComputer Use Agents, GUI Agents, Computer-Using Agent (CUA)
Using AI
A capability that lets an AI model operate a real desktop or browser the way a person would: by looking at screenshots…
Computer VisionCV, Machine Vision
Fundamentals
The field of AI focused on enabling machines to interpret and understand visual information from the world &mdash…
Constitutional AICAI
safety
An alignment technique developed by Anthropic where a model is trained to follow a set of principles (a "constitution")…
ContaminationData Contamination, Benchmark Leaking
fundamentals
When benchmark test data appears in a model's training data, inflating its scores without reflecting genuine…
Content ModerationAI Moderation, Trust & Safety
Safety
Using AI to detect and filter harmful, illegal, or policy-violating content at scale. This includes text classification…
Context Engineering
Using AI
The practice of designing, assembling, and maintaining everything a model sees in its context window — the system…
Context Length ExtensionYaRN, NTK Scaling, RoPE Scaling
infra
Techniques that enable language models to handle sequences longer than those seen during training. A model trained on…
Context WindowContext Length
Using AI
The maximum amount of text (measured in tokens) a model can process in a single conversation. This includes both your…
Continual LearningLifelong Learning, Incremental Learning
training
The ability of a model to learn from new data continuously without forgetting what it learned before. Current LLMs are…
Contrastive LearningSimCLR, InfoNCE
training
A self-supervised learning approach that trains models by contrasting positive pairs (similar items that should be…
ControlNet
models
An architecture that adds spatial control to image generation models. Instead of just describing what you want in text…
ConvolutionConv, Convolutional Layer, Kernel, Filter
fundamentals
A mathematical operation that slides a small filter (kernel) across an input to detect local patterns. In images, a…
CoreWeave
companies
A specialized cloud provider built entirely around GPU computing for AI workloads. CoreWeave operates large clusters of…
CorpusDataset, Training Data
Training
The body of text (or other data) used to train a model. A corpus can range from curated collections of books and papers…
Cosine SimilarityCosine Distance, Vector Similarity
fundamentals
A measure of similarity between two vectors based on the angle between them, ignoring their magnitude. Cosine…
Cross-AttentionEncoder-Decoder Attention
fundamentals
An attention mechanism where the queries come from one sequence and the keys/values come from a different sequence. In…
Cross-ValidationK-Fold CV, Leave-One-Out
training
A technique for evaluating model performance when you don't have enough data for a separate test set. K-fold…
Curriculum Learning
training
A training strategy that presents examples in a meaningful order — typically from easy to hard — rather…
Cursor
companies
An AI-native code editor built as a fork of VS Code, integrating LLMs deeply into the editing experience: inline code…
DALL-EDALL-E 2, DALL-E 3
models
OpenAI's image generation model family. DALL-E 1 (2021) used a discrete VAE + Transformer approach. DALL-E 2 (2022)…
Data Augmentation
training
Techniques that artificially expand a training dataset by creating modified versions of existing examples. For images…
Data CentersAI Data Centers, GPU Clusters
Infrastructure
Physical facilities that house the servers, GPUs, networking equipment, and cooling systems needed to train and run AI…
DatabricksMosaic ML, DBRX, Unity Catalog
companies
A data and AI platform that provides unified analytics, data engineering, and machine learning capabilities. Databricks…
DatasetTraining Set, Data
fundamentals
A structured collection of data used to train, evaluate, or test a machine learning model. Datasets can be labeled…
Decart AIReal-time world simulation, game generation
Companies
Israeli AI company pushing the boundaries of real-time AI generation. Their technology can generate interactive…
DecoderDecoder Network, Generator
fundamentals
A neural network component that generates output from a representation. In Transformers, the decoder uses causal…
Deep LearningDL
Fundamentals
A subset of machine learning that uses neural networks with many layers (hence "deep") to learn hierarchical…
Deep ResearchDeep Research Agents, Agentic Research
Using AI
A mode offered by several AI assistants where the model autonomously plans and carries out a multi-step web research…
DeepfakesSynthetic Media, AI-Generated Fakes
Safety
AI-generated images, video, or audio designed to convincingly depict real people saying or doing things they never did.…
DeepgramNova speech-to-text, Aura text-to-speech
Companies
Speech AI company building fast, accurate speech recognition and text-to-speech APIs. Their Nova models compete with…
DeepLNeural machine translation, DeepL Pro
Companies
German AI company widely regarded as the best machine translation service in the world. Built by a team of…
DeepSeekDeepSeek-V3, DeepSeek-R1
Companies
Chinese AI lab that shook the industry in early 2025 with DeepSeek-R1, a reasoning model rivaling frontier labs at a…
Developer ToolsAI SDKs, AI Frameworks
Tools
The ecosystem of libraries, frameworks, and platforms that make building AI-powered applications easier. This includes…
DevinCognition Devin, Cognition AI
Tools
An autonomous AI software engineer built by Cognition AI that takes high-level tasks and carries them out end to end…
Differential PrivacyDP
safety
A mathematical framework that guarantees individual privacy in aggregate data analysis and model training. With…
Diffusion LLMText Diffusion Model, dLLM
Models
A large language model that generates text through iterative denoising instead of predicting one token at a time from…
Diffusion Model
Models
A type of generative model that creates images (or video, audio) by starting with pure noise and gradually removing it…
Diffusion TransformerDiT
models
An architecture that replaces the U-Net backbone traditionally used in diffusion models with a Transformer. DiT applies…
DistillationKnowledge Distillation
training
Training a smaller "student" model to mimic a larger "teacher" model by learning from the teacher's soft probability…
Distributed TrainingData Parallelism, Model Parallelism, FSDP
infra
Training a model across multiple GPUs or machines simultaneously. Data parallelism gives each GPU a copy of the model…
DPODirect Preference Optimization
training
An alternative to RLHF for aligning language models with human preferences. DPO directly optimizes the model using…
Drift DetectionData Drift, Model Drift, Concept Drift
infra
Monitoring for changes in the data distribution or model behavior over time that could degrade performance. Data drift…
DropoutRegularization, Weight Decay
training
A regularization technique that randomly "turns off" a fraction of neurons during each training step by setting their…
Dual UseDual-Use Technology
safety
Technology that can be used for both beneficial and harmful purposes. AI is inherently dual-use: the same model that…
Early StoppingPatience, Validation-Based Stopping
training
Stopping training when performance on a held-out validation set stops improving, rather than training for a fixed…
Edge AIOn-Device AI
infra
Running AI models directly on end-user devices — phones, laptops, cars — rather than in the cloud. Your…
ElevenLabsVoice synthesis, voice cloning, dubbing
Companies
Voice AI company that made ultra-realistic speech synthesis accessible to everyone. Their technology powers voice…
EmbeddingVector Embedding
Training
A way to represent text (or images, or audio) as a list of numbers (a vector) that captures its meaning. Similar…
Embedding LayerToken Embedding, Embedding Table, Lookup Table
fundamentals
A lookup table that maps each token in the vocabulary to a dense vector (the token's embedding). When the model…
Embodied AIEmbodied Intelligence, Robotics Foundation Models
Fundamentals
An approach to artificial intelligence where models control physical machines — robots that perceive, move, and…
EmergenceEmergent Abilities, Emergent Behavior
Fundamentals
Capabilities that appear in AI models at scale but were not explicitly trained for — abilities that seem to…
EncoderEncoder Network, Feature Extractor
fundamentals
A neural network component that converts input data into a compressed, information-rich representation (encoding). In…
Encoder-DecoderSeq2Seq
models
A model architecture with an encoder that compresses input and a decoder that generates output from it. T5 and BART are…
Endpoint
Infrastructure
A specific URL where an AI API accepts requests. For example, Anthropic's message endpoint is where you send prompts to…
EU AI ActAI Act, European AI Act
Safety
The European Union's comprehensive law regulating artificial intelligence, in force since August 2024. It takes a…
EvaluationEvals, Model Evaluation
Training
The methods used to measure how well an AI model performs. This goes far beyond benchmarks — it includes human…
Existential RiskX-Risk, AI Doom
safety
The hypothesis that sufficiently advanced AI systems could pose a threat to human existence or permanently curtail…
Facial RecognitionFace Recognition, Face ID
safety
Identifying or verifying a person from their face in an image or video. Verification asks "is this person who they…
FalconTII Falcon, Falcon LLM
Models
An open-weight family of large language models built by the Technology Innovation Institute (TII), the applied research…
FeatureLearned Representation, Activation
fundamentals
A pattern or concept that a neural network learns to detect in its input. In vision, early-layer features are edges and…
Federated LearningFL, Collaborative Learning
training
A training approach where the model is trained across multiple devices or organizations without sharing the raw data.…
Feedforward NetworkFFN, MLP Block
fundamentals
The component in each Transformer layer that processes each token independently through two linear transformations with…
Few-Shot LearningIn-Context Learning
using
Providing example input-output pairs in your prompt to teach the model a pattern. Zero-shot = no examples, one-shot =…
Fine-tuning
Training
Taking a pre-trained model and training it further on a smaller, specific dataset to specialize its behavior. Like…
Fireworks AIFireworks, fireworks.ai
Companies
Fireworks AI is an inference platform that serves open-weight AI models behind fast, pay-per-use APIs. Founded in 2022…
Flash AttentionFlashAttention, FlashAttention-2
infra
A GPU-optimized implementation of the attention mechanism that is 2–4x faster and uses significantly less memory…
FLOPsFloating Point Operations, FLOP/s, Compute
fundamentals
Floating Point Operations — the standard measure of computational work in AI. Training a model requires a certain…
Flow MatchingRectified Flow
models
A generative technique that transforms noise into data by following smooth, direct paths. Fewer steps than diffusion…
Foundation Model
Fundamentals
A large model trained on broad data that serves as a base for many different tasks. Claude, GPT, Gemini, and Llama are…
Function CallingTool Calling, Tool Use API
using
A structured way for AI models to request execution of external functions during a conversation. You define functions…
GANGenerative Adversarial Network
Models
A model architecture where two neural networks compete: a generator creates fake data, and a discriminator tries to…
GeminiGoogle Gemini, Bard
Models
Google's flagship family of large language models, built by Google DeepMind as the successor to PaLM and the Bard…
GemmaGoogle Gemma, Gemma 3
Models
Google's family of open-weight language models, created by Google DeepMind from the same research and technology behind…
Generative AIGenAI
Fundamentals
AI systems that create new content — text, images, audio, video, code, 3D models — rather than just…
GGUFGGML Unified Format
infra
The standard file format for running quantized language models locally via llama.cpp, Ollama, and other local inference…
GitHub CopilotCopilot
Tools
An AI coding assistant built by GitHub in partnership with OpenAI, launched in 2021 and widely credited as the first…
GNNGraph Neural Network
models
Neural networks designed to operate on graph-structured data — data where entities are connected by relationships…
Google DeepMindGemini, AlphaGo, AlphaFold
Companies
Google's unified AI research division, formed by merging DeepMind and Google Brain in 2023. Behind Gemini, AlphaGo…
GPTChatGPT, GPT-4, GPT-5
Models
GPT (Generative Pre-trained Transformer) is OpenAI's family of large language models: decoder-only transformers trained…
GPUGraphics Processing Unit
Infrastructure
Originally designed for rendering graphics, GPUs turned out to be perfect for AI because they can do thousands of math…
GQAGrouped Query Attention
fundamentals
An attention variant where multiple query heads share a single key-value head, reducing the KV cache size without…
Gradient CheckpointingActivation Checkpointing, Rematerialization
training
A memory-saving technique that trades compute for memory during training. Instead of storing all intermediate…
Gradient DescentSGD, Backpropagation
training
The algorithm that trains neural networks by iteratively adjusting parameters to reduce the loss. Computes how much…
GrokxAI Grok
Models
Grok is the family of large language models built by xAI, Elon Musk's artificial intelligence company, and the name of…
GroqGroq LPU
companies
A chip company building custom AI inference processors (LPUs) purpose-built for sequential token generation, achieving…
Grounding
Using AI
Connecting a model's responses to factual, verifiable sources rather than letting it rely solely on its training data.…
GRPOGroup Relative Policy Optimization
Training
A reinforcement learning algorithm for fine-tuning language models, introduced by DeepSeek in 2024 and popularized by…
Guardrails
Safety
Safety mechanisms that prevent AI models from generating harmful, inappropriate, or off-topic content. Guardrails can…
Guidance ScaleCFG Scale, Classifier-Free Guidance
using
A parameter that controls how strongly an image generation model follows the text prompt. Low guidance (1–3): the…
Hallucination
Using AI
When an AI model generates information that sounds confident and plausible but is factually wrong or entirely…
HeyGenAI avatar videos, lip-sync dubbing
Companies
AI video platform specializing in realistic talking-head avatars and automatic lip-sync dubbing. Used by enterprises…
HiDreamHiDream image generation models
Companies
Emerging image generation company building high-quality diffusion models. Their open-weights releases have gained…
Hugging FaceHF
companies
The central hub of open-source AI. Hosts 500K+ models, 100K+ datasets, the Transformers library, and Spaces for demos.…
Human EvaluationHuman Eval, Manual Evaluation
fundamentals
Evaluating AI output quality by having humans judge it directly. Humans assess fluency, accuracy, helpfulness, safety…
HumeEmpathic Voice Interface, emotion detection
Companies
AI company building models that understand and express human emotion. Their Empathic Voice Interface detects tone…
Hyperparameter TuningHPO, Hyperparameter Optimization, Grid Search
training
Systematically searching for the best hyperparameters — the configuration choices that aren't learned during…
HyperparametersTraining Hyperparameters
Training
Settings you choose before training begins that control how the model learns — as opposed to parameters, which…
IdeogramText rendering in images, Ideogram 2.0
Companies
AI image generation company founded by former Google Brain researchers. Made their name by solving one of the hardest…
Image GenerationText-to-Image, AI Art
fundamentals
Creating images from text descriptions using AI models. You type "a sunset over mountains in watercolor style" and the…
Image SegmentationSemantic Segmentation, SAM, Instance Segmentation
using
Classifying every pixel in an image into a category. Semantic segmentation labels pixels by class (road, sidewalk…
Image-to-Imageimg2img, Image Conditioning
using
Generating a new image based on an existing image plus a text prompt. Instead of starting from pure noise…
ImageNetImageNet Challenge, ILSVRC
Training
A large-scale image dataset of roughly 14 million hand-labeled images organized into more than 20,000 categories, first…
In-Context LearningICL
Fundamentals
The ability of a large language model to pick up a task from examples or instructions placed in the prompt, with no…
Induction Head
fundamentals
A specific two-attention-head circuit discovered in Transformers that implements in-context learning by pattern…
Inference
Infrastructure
The process of running a trained model to generate outputs. Training is learning; inference is using what was learned.…
Inflection AIInflection, Pi
Companies
An AI company founded in 2022 by Mustafa Suleyman, Reid Hoffman, and Karén Simonyan, best known for Pi, a…
Information ExtractionIE, Structured Extraction
using
Automatically extracting structured information from unstructured text. Given a news article, extract: who did what…
InpaintingImage Inpainting, Outpainting
using
Filling in a selected region of an image with AI-generated content that matches the surrounding context. You mask an…
Instruction FollowingInstruction Adherence
using
A model's ability to accurately execute what the user asks for — respecting format constraints, length…
Instruction TuningInstruction Fine-Tuning, IFT, SFT
training
Fine-tuning a pre-trained language model on a dataset of (instruction, response) pairs to teach it to follow…
JailbreakJailbreaking, Adversarial Prompt
safety
Techniques that trick an AI model into bypassing its safety training and generating content it was designed to refuse…
Jina AIEmbeddings, Reader API, rerankers
Companies
Berlin-based AI company specializing in search and embeddings. Their jina-embeddings models and Reader API (which…
Kling AIKling video generation, long-form video
Companies
AI video platform from Kuaishou (China's second-largest short-video platform). Gained rapid international attention for…
Knowledge CutoffTraining Data Cutoff, Knowledge Date
fundamentals
The date after which a model has no training data, meaning it lacks knowledge of events, discoveries, or changes that…
Knowledge EditingModel Editing, Fact Editing
training
Techniques for modifying specific facts in a trained model without retraining it. If a model incorrectly states "The…
Knowledge GraphKG, Ontology
fundamentals
A structured representation of knowledge as a network of entities (nodes) connected by relationships (edges). "Paris…
KV CacheKey-Value Cache
infra
A memory optimization storing previously computed attention key/value tensors so they don't need recomputation for each…
Lambda LabsLambda, Lambda Cloud
companies
A GPU cloud provider focused specifically on AI and machine learning workloads. Lambda offers on-demand and reserved…
LangChain
tools
A popular open-source framework for building applications with language models. LangChain provides abstractions for…
Language DetectionLanguage Identification, LangID
using
Automatically identifying which language a text is written in. "Bonjour le monde" → French. "こんにちは世界" &rarr…
Large Language ModelLLM
Fundamentals
A neural network trained on massive amounts of text to understand and generate human language. "Large" refers to the…
LatencyTime to First Token (TTFT)
Infrastructure
The delay between sending a request and getting the first response. In AI, this is often measured as Time to First…
LayerHidden Layer, Neural Network Layer
fundamentals
A group of neurons that processes data at a specific level of abstraction in a neural network. The input layer receives…
Learning Rate ScheduleLR Schedule, Warmup, Cosine Annealing
training
A strategy for changing the learning rate during training rather than keeping it constant. Most modern training uses…
Leonardo.aiCreative image generation, game asset creation
Companies
Australian AI image platform that carved out a niche between Midjourney and Stable Diffusion. Popular with game…
Liquid AILiquid Foundation Models, liquid neural networks
Companies
MIT spinout exploring fundamentally different neural network architectures inspired by biological neural circuits.…
LlamaLLaMA, Meta Llama
Models
A family of large language models from Meta whose weights are freely downloadable under a custom community license.…
llama.cpp
tools
An open-source C/C++ library for running LLM inference on consumer hardware, created by Georgi Gerganov. llama.cpp…
LlamaIndexGPT Index
Tools
An open-source data framework for building LLM applications over private data. LlamaIndex handles the plumbing between…
LLM-as-JudgeLLM Judge, LLM-as-a-Judge
Training
An evaluation technique where a large language model grades the outputs of another model (or its own) instead of…
LogitsRaw Scores, Pre-Softmax Outputs
fundamentals
The raw, unnormalized scores that a model outputs before they're converted into probabilities by the softmax function.…
LoRALow-Rank Adaptation
Training
A technique that makes fine-tuning dramatically cheaper by only training a small number of additional parameters…
Loss FunctionObjective Function
training
A mathematical function measuring how wrong a model's predictions are. For LLMs, cross-entropy loss measures how…
LSTMLong Short-Term Memory
models
A type of recurrent neural network (RNN) designed to learn long-range dependencies in sequential data. LSTM introduces…
Luma AIDream Machine, Ray2
Companies
AI company focused on video and 3D generation. Their Dream Machine was one of the first accessible, high-quality AI…
Machine LearningML
Fundamentals
The broad field of computer science where systems learn patterns from data rather than following explicit rules.…
Machine TranslationMT, Neural Machine Translation, NMT
using
Automatically translating text from one language to another. Modern neural machine translation (NMT) uses…
MambaSelective SSM
models
A selective state space model architecture challenging the Transformer. Achieves competitive performance with linear…
Masked Language ModelingMLM, Masked LM, Cloze Task
training
A self-supervised training objective where random tokens in the input are replaced with a [MASK] token, and the model…
Matrix MultiplicationMatmul, GEMM
fundamentals
The fundamental mathematical operation underlying all neural networks. Multiplying a weight matrix by an input vector…
MCPModel Context Protocol
Tools
An open protocol (created by Anthropic) that standardizes how AI models connect to external tools and data sources.…
Mechanistic InterpretabilityMech Interp
safety
Reverse-engineering what happens inside neural networks at the level of neurons, circuits, and features — not…
MemoryAI Memory, Persistent Context
Using AI
Mechanisms that allow AI models to retain and recall information beyond a single conversation. This includes in-context…
Meta AILlama, FAIR, PyTorch
Companies
Meta's AI research division, home of FAIR (Fundamental AI Research). Responsible for the open-weights Llama model…
Midjourney
companies
An AI image generation company known for aesthetically refined output. Operates through Discord and web. Runs…
MiniMaxMiniMax models, Hailuo AI, video generation
Companies
Chinese AI company building large-scale models across text, voice, and video. Known for their Hailuo consumer platform…
Mistral AIMistral, Mixtral, Codestral, Le Chat
Companies
European AI powerhouse founded by former DeepMind and Meta researchers. Known for punching above their weight with…
Mixed Precision TrainingFP16, BF16, Half Precision
training
Training neural networks using lower-precision number formats (16-bit instead of 32-bit) for most computations while…
Mixture of ExpertsMoE
Models
An architecture where the model contains multiple "expert" sub-networks, but only activates a few of them for each…
ModelAI Model, ML Model
Fundamentals
A trained mathematical system that takes inputs and produces outputs based on patterns learned from data. In AI…
Model CardModel Documentation, Data Sheet
safety
A standardized document that describes a machine learning model's intended use, performance characteristics, training…
Model CollapseData Feedback Loop
training
The degradation that occurs when AI models are trained on data generated by previous AI models, creating a feedback…
Model MergingTIES, DARE, SLERP, Frankenmerge
training
Combining the weights of multiple fine-tuned models into a single model without any additional training. If model A is…
Model RegistryModel Store, Model Catalog
infra
A centralized system for versioning, tracking, and managing trained machine learning models throughout their lifecycle.…
Model ServingvLLM, TGI, TensorRT-LLM, Inference Server
infra
The infrastructure and software that runs trained AI models in production, handling incoming requests, managing GPU…
Moonshot AIKimi, ultra-long context models
Companies
Chinese AI company that made waves by launching Kimi, a chatbot with a 2-million-token context window. Founded by Yang…
Multi-Agent SystemsMulti-Agent, Agent Swarm
using
Architectures where multiple AI agents collaborate, debate, or specialize to solve problems that a single agent can't…
Multi-Head AttentionMHA
fundamentals
Running multiple attention operations in parallel, each with its own learned projection of the queries, keys, and…
Multimodal
Fundamentals
A model that can understand and/or generate multiple types of data: text, images, audio, video, code. Claude can read…
Music GenerationAI Music, Text-to-Music
using
Creating music from text descriptions, melodies, or other audio inputs using AI models. "An upbeat electronic track…
Named Entity RecognitionNER, Entity Extraction
using
Identifying and categorizing named entities in text — people, organizations, locations, dates, monetary amounts…
Natural Language ProcessingNLP
Fundamentals
The branch of AI focused on enabling machines to understand, interpret, and generate human language. NLP covers…
Negative PromptNegative Conditioning
using
A text description of what you don't want in a generated image, used alongside the main prompt. Prompt: "a beautiful…
Neural Network
Fundamentals
A computing system loosely inspired by biological brains, made of layers of interconnected "neurons" (mathematical…
NeuronArtificial Neuron, Perceptron, Node
fundamentals
The basic computational unit of a neural network. An artificial neuron receives inputs, multiplies each by a weight…
NormalizationLayerNorm, RMSNorm, BatchNorm
training
Techniques that stabilize neural network training by normalizing the values flowing through the network to have…
NVIDIAGPUs, CUDA, H100/H200, NeMo
Companies
The company whose GPUs power virtually all AI training and most inference worldwide. What started as a graphics card…
Object DetectionYOLO, Bounding Box Detection
using
Identifying and localizing objects in images or video by drawing bounding boxes around them and classifying what each…
OCROptical Character Recognition, Text Recognition
using
Extracting text from images — photographs of documents, screenshots, signs, handwritten notes, or any image…
Ollama
tools
A user-friendly tool for running language models locally with a single command. Ollama wraps llama.cpp in a Docker-like…
ONNXOpen Neural Network Exchange
infra
An open format for representing machine learning models that enables interoperability between frameworks. A model…
Open vs. ClosedOpen Source vs. Proprietary, Open Weights Debate
fundamentals
The ongoing debate about whether AI models should be openly released (weights publicly available, like Llama and…
Open WeightsOpen Source (in AI context)
Safety
When a company releases a model's trained parameters for anyone to download and run. "Open weights" is more accurate…
OpenAIGPT, ChatGPT, DALL-E, Sora
Companies
The company behind ChatGPT and the GPT series of models. Originally a non-profit research lab, OpenAI became the public…
OpenRouter
Companies
A unified API gateway that gives developers access to hundreds of large language models from different providers…
OptimizationModel Optimization, Inference Optimization
Training
The broad set of techniques used to make AI models faster, smaller, cheaper, or more accurate. This includes training…
Overfitting
Training
When a model memorizes its training data too well and loses the ability to generalize to new inputs. Like a student who…
PagedAttention
infra
A memory management technique for KV cache that borrows from operating system virtual memory. Instead of allocating a…
ParametersWeights, Model Parameters
Fundamentals
The internal values a neural network learns during training — essentially the "knowledge" of the model encoded as…
PerplexityAI-powered search engine, Sonar API
Companies
AI search engine that combines real-time web search with language model reasoning to give direct, sourced answers…
Perplexity (Metric)PPL
fundamentals
A measurement of how well a language model predicts text. Represents how many tokens the model is choosing between at…
PhiMicrosoft Phi
Models
Microsoft's family of small language models, ranging from roughly 1 billion to 14 billion parameters, built to deliver…
PikaPika Labs
Companies
An AI video generation company whose web app turns text prompts, still images, and uploaded footage into short video…
PixVersePixVerse video generation
Companies
Chinese video generation company building accessible AI video tools. Known for fast generation speeds and a free tier…
PoolingMax Pooling, Average Pooling
fundamentals
An operation that reduces the spatial dimensions of data by summarizing a region into a single value. Max pooling takes…
Pose EstimationBody Pose, Skeleton Detection, Keypoint Detection
using
Detecting the position and orientation of a human body (or animal, hand, face) in an image or video by locating key…
Positional EncodingPositional Embedding, RoPE, ALiBi
fundamentals
A mechanism that tells a Transformer model the order of tokens in a sequence. Unlike RNNs which process tokens…
Post-TrainingPost-training
Training
The set of training stages applied to a model after pretraining, turning a raw next-token predictor into a usable…
PPOProximal Policy Optimization
Training
PPO (Proximal Policy Optimization) is a reinforcement learning algorithm that updates a policy in small, controlled…
Pre-training
Training
The initial, massive training phase where a model learns language (or other modalities) from a huge corpus. This is the…
Precision & RecallF1 Score, Confusion Matrix
fundamentals
Two complementary metrics for evaluating classifiers. Precision answers "of the items the model flagged as positive…
Prompt
fundamentals
The text you give to an AI model to get a response. A prompt can be a question, an instruction, a creative brief, or…
Prompt CachingContext Caching, Prefix Caching
using
A technique that saves and reuses the processed version of a prompt prefix across multiple API calls, avoiding…
Prompt Engineering
Using AI
The practice of crafting inputs to get better outputs from AI models. This ranges from simple techniques (being…
Prompt InjectionIndirect Prompt Injection
safety
An attack where malicious instructions are embedded in content that an AI model processes, causing the model to follow…
Prompt TemplateTemplate, Prompt Pattern
using
A reusable prompt structure with variable placeholders that gets filled in with specific data at runtime. Instead of…
PruningModel Pruning, Weight Pruning
training
Removing unnecessary parameters (weights, neurons, or entire layers) from a trained model to make it smaller and faster…
QuantizationGGUF, GPTQ, AWQ
Infrastructure
Reducing a model's precision to make it smaller and faster. A model trained in 32-bit floating point can be quantized…
Question AnsweringQA, Reading Comprehension
using
A system that answers questions posed in natural language. Extractive QA finds the answer span within a given document…
QwenTongyi Qianwen
Models
Qwen is a family of open-weight large language models developed by Alibaba Cloud. The lineup spans several generations…
RAGRetrieval-Augmented Generation
Tools
A technique that gives AI models access to external knowledge by retrieving relevant documents before generating a…
Rate Limiting
Infrastructure
Restrictions on how many API requests you can make per minute/hour/day. Providers impose rate limits to prevent server…
ReasoningAI Reasoning, Chain-of-Thought Reasoning
Using AI
The ability of AI models to think step-by-step, decompose complex problems, and arrive at logically sound conclusions.…
RecraftRecraft V3, vector graphics generation
Companies
AI design tool focused on professional-grade image and vector graphic generation. One of the first to produce truly…
Red Teaming
Safety
The practice of deliberately trying to make an AI model fail, misbehave, or produce harmful outputs. Red teams probe…
RegressionLinear Regression, Prediction
fundamentals
A machine learning task that predicts a continuous numerical value rather than a category. "What will the temperature…
Reinforcement LearningRL
Training
A training paradigm where an AI agent learns by interacting with an environment, taking actions, and receiving rewards…
RekaReka Core, Reka Flash
Companies
AI research company founded by former DeepMind, Google Brain, and FAIR researchers. Building natively multimodal models…
Replicate
Companies
A cloud platform for running machine learning models through a simple API, with no infrastructure to manage. It hosts…
RerankingRe-ranking, Cross-Encoder Reranking
Tools
A second scoring pass in a retrieval pipeline that re-orders an initial set of candidate documents with a more accurate…
Resemble AIVoice cloning, speech synthesis, watermarking
Companies
Canadian voice AI company specializing in high-fidelity voice cloning and real-time speech synthesis. One of the first…
Residual ConnectionSkip Connection, Shortcut Connection
fundamentals
A connection that bypasses one or more layers by adding the input directly to the output: output = layer(x) + x.…
RetrievalInformation Retrieval, IR
fundamentals
The process of finding relevant documents, passages, or data from a large collection in response to a query. In AI…
Reward ModelRM, Preference Model
training
A model trained to predict human preferences between AI responses. Given a prompt and two candidate responses, the…
RLAIFRL from AI Feedback
training
A variant of RLHF where the preference labels come from an AI model instead of human annotators. A strong AI model…
RLHFReinforcement Learning from Human Feedback
Training
A training technique where human evaluators rank model outputs by quality, and this feedback is used to train a reward…
RLVRReinforcement Learning with Verifiable Rewards, RL with Verifiable Rewards
Training
A reinforcement learning method in which the reward signal comes from automatic, programmatic checkers — such as…
RNNRecurrent Neural Network, LSTM, GRU
models
A neural network that processes sequences by maintaining a hidden state that gets updated at each step — it…
RunwayGen-1, Gen-2, Gen-3 Alpha
Companies
Pioneering AI video generation company. Co-created the original Stable Diffusion architecture and then pivoted to…
Safe SuperintelligenceSSI
Companies
Safe Superintelligence (SSI) is an AI research company founded in 2024 by Ilya Sutskever, Daniel Gross, and Daniel…
SambaNovaSN40L chip, ultra-fast inference
Companies
AI hardware company that designs custom chips (RDUs) purpose-built for AI workloads. Their SambaNova Cloud offers some…
SamplingDecoding Strategy, Top-p, Top-k
fundamentals
The process of selecting which token to generate next from the model's predicted probability distribution. Greedy…
Sarvam AISarvam models, Indian language AI
Companies
Indian AI company building models specifically optimized for India's linguistic diversity. Their models handle Hindi…
Scale AI
companies
The largest AI data labeling company, providing the human-annotated training data that most major AI models rely on.…
Scaling LawsChinchilla Scaling
fundamentals
Empirical power-law relationships: model performance improves predictably with more parameters, data, and compute. You…
Self-AttentionScaled Dot-Product Attention
fundamentals
An attention mechanism where a sequence attends to itself — every token computes its relevance to every other…
Self-Supervised LearningSSL
training
Training where the model generates its own supervision from unlabeled data by hiding part of the input and predicting…
Sentiment AnalysisOpinion Mining
using
Automatically determining the emotional tone of text — positive, negative, or neutral. "This product is amazing!"…
SGLangSGLang Runtime, SRT
Tools
SGLang is an open-source serving engine for large language models, built to run inference fast and cheaply at…
SigmoidLogistic Function
fundamentals
A mathematical function that squashes any real number into the range (0, 1): σ(x) = 1 / (1 + e^(−x)).…
SLMSmall Language Model, Small Language Models
Fundamentals
A language model designed for efficiency, typically ranging from about 100 million to 15 billion parameters. SLMs are…
SlopAI Slop, Generated Slop
Safety
Low-quality, generic, unwanted AI-generated content that floods the internet. The term emerged in 2024 as a pejorative…
SoftmaxSoftmax Function, Normalized Exponentials
fundamentals
A function that converts a vector of raw numbers (logits) into a probability distribution — all values become…
SoraOpenAI Sora
Models
Sora is OpenAI's video generation model family, which creates short video clips from text prompts or still images.…
Sovereign AISovereign AI
Infrastructure
Sovereign AI is a national strategy of building domestic artificial-intelligence capability — a country's own…
Sparse AttentionLocal Attention, Sliding Window Attention
models
Attention mechanisms that process only a subset of token pairs instead of the full N×N attention matrix. Sliding…
Sparse AutoencoderSAE
models
A neural network trained to reconstruct a model's internal activations through a bottleneck with a sparsity constraint…
Speaker DiarizationWho Spoke When
using
Determining who spoke when in an audio recording with multiple speakers. Given a meeting recording, diarization…
Speculative DecodingAssisted Generation
infra
A small draft model generates candidate tokens, then the large model verifies them all at once. Correct guesses (common…
Speech RecognitionSTT, Speech-to-Text, ASR
using
Converting spoken audio into text. Modern speech recognition uses deep learning models (most notably OpenAI's Whisper)…
Stability AIStable Diffusion, SDXL, Stable Audio
Companies
The company that democratized image generation by releasing Stable Diffusion as open-source in 2022. Despite leadership…
Stable DiffusionSD, SDXL, SD3
models
The most widely used open-source image generation model, created by Stability AI in collaboration with academic…
State Space ModelSSM, Mamba
Models
An alternative to Transformers that processes sequences by maintaining a compressed "state" instead of using attention…
StepFunStep models, multimodal AI
Companies
Chinese AI startup building competitive large language and multimodal models. Their Step series has shown strong…
Stochastic Parrot
Safety
A critique of large language models arguing that they are merely sophisticated pattern matchers that stitch together…
StreamingToken Streaming
using
Sending model output token by token as generated, via Server-Sent Events. This is why chat interfaces show text…
Structured OutputJSON Mode
using
Getting AI to respond in machine-parseable format like JSON. Most providers support this natively: define a schema, the…
Style TransferNeural Style Transfer
using
Applying the visual style of one image (a painting, a photograph, a design) to the content of another image. "Make this…
SunoAI music generation
Companies
AI music generation company that lets anyone create full songs — vocals, instruments, production — from a…
Super ResolutionUpscaling, Image Enhancement, SR
using
Increasing the resolution of an image by generating plausible detail that wasn't in the original. A 256×256 photo…
SuperpositionFeature Superposition, Polysemanticity
fundamentals
The phenomenon where neural networks encode many more features (concepts, patterns) than they have neurons, by…
Supervised Learning
training
Training from labeled examples where the correct answer is provided. The model adjusts to minimize the difference…
SwiGLUGated Linear Unit, GLU Variants
fundamentals
A gated activation function used in the feedforward layers of modern Transformers. SwiGLU combines the SiLU/Swish…
SycophancyAI Sycophancy, People-Pleasing
Safety
The tendency of AI models to tell users what they want to hear rather than what's true. A sycophantic model agrees with…
Synthetic DataAI-Generated Training Data
training
Training data generated by AI models rather than collected from real sources. A frontier model generates examples used…
System PromptSystem Message
Using AI
A special instruction given to a model at the start of a conversation that sets its behavior, personality, and rules.…
Temperature
Using AI
A parameter that controls how random or deterministic a model's output is. Temperature 0 makes the model always pick…
TencentHunyuan, WeChat, gaming AI
Companies
Chinese tech giant behind WeChat, one of the world's largest gaming companies, and increasingly a force in generative…
TensorMultidimensional Array
fundamentals
A multidimensional array of numbers — the fundamental data structure in deep learning. A scalar is a 0D tensor (a…
Test-Time ComputeInference-Time Compute, Chain of Thought, Thinking Tokens
fundamentals
Using additional computation during inference (when the model is generating a response) to improve answer quality.…
Text SummarizationSummarization, TL;DR
using
Automatically generating a shorter version of a text that preserves the key information. Extractive summarization…
Text-to-SpeechTTS, Speech Synthesis, Voice AI
using
Converting written text into natural-sounding spoken audio. Modern TTS systems use neural networks to generate speech…
Thinking MachinesThinking Machines Lab, TML
Companies
An AI research and product company founded in 2025 by former OpenAI CTO Mira Murati and a team of prominent…
ThroughputTokens Per Second, TPS
infra
The total number of tokens a system can generate per second across all concurrent requests. Distinct from latency (how…
Together AI
companies
A cloud platform for running and training open-source AI models. Together AI provides inference APIs for popular open…
Token
Fundamentals
The basic unit of text that AI models process. A token is typically a word or word fragment — "understanding"…
TokenizerTokenization
fundamentals
The algorithm converting raw text into tokens before the model sees it. Different models use different tokenizers…
Tool UseFunction Calling
Tools
The ability of an AI model to call external functions or tools during a conversation. Instead of just generating text…
TPUTensor Processing Unit
Infrastructure
A family of custom AI accelerator chips (ASICs) designed by Google specifically for machine learning workloads. TPUs…
Transfer Learning
training
Using knowledge learned from one task or dataset to improve performance on a different but related task. Instead of…
Transformer
Models
The neural network architecture behind virtually all modern LLMs and many image/audio models. Introduced by Google in…
TripoText-to-3D, image-to-3D generation
Companies
AI company specializing in generating 3D models from text or images. In a field where most 3D generation produces…
Turing TestImitation Game
Fundamentals
A behavioral test for machine intelligence proposed by Alan Turing in his 1950 paper “Computing Machinery and…
Twelve LabsVideo search, Pegasus, Marengo
Companies
Video understanding company that lets you search, analyze, and generate content from video using natural language.…
Udio
Companies
Udio is an AI music generation company whose web app turns text prompts into complete songs, including vocals, lyrics…
Unsupervised Learning
training
Finding patterns in data without labels. Clustering, dimensionality reduction, and anomaly detection are classic tasks.…
UpstageSolar models, Document AI
Companies
Korean AI company known for their Solar model family and Document AI products. Demonstrated that smaller, well-trained…
Validation SetDev Set, Hold-Out Set
training
A subset of data held back from training, used to evaluate model performance during development and tune…
Vector DatabaseQdrant, Pinecone, Weaviate, ChromaDB
Tools
A database optimized for storing and searching embeddings (vectors). Instead of matching exact keywords like a…
VeoGoogle Veo, Veo 2, Veo 3
Models
Google DeepMind's family of video generation models, which turn text prompts or still images into short…
Vibe CodingVibe Coding
Using AI
Vibe coding is a way of building software where you describe what you want in natural language, let an AI model write…
Video GenerationText-to-Video, AI Video
fundamentals
Creating video from text descriptions, images, or other videos using AI models. Sora (OpenAI), Kling (Kuaishou), Runway…
ViduVidu video generation, long-form coherent video
Companies
Video generation platform from Shengshu Technology, producing some of the most physically coherent AI-generated videos.…
VisionMultimodal Vision, Image Understanding
using
The ability of a language model to understand and reason about images alongside text. You send a photo and ask "what's…
Vision TransformerViT
models
A Transformer architecture applied to images by splitting an image into fixed-size patches (e.g., 16×16 pixels)…
vLLM
tools
An open-source LLM serving engine that achieves high throughput through PagedAttention and continuous batching. vLLM…
VocabularyVocab, Token Vocabulary
fundamentals
The fixed set of tokens that a model can recognize and produce. A vocabulary is built by the tokenizer during training…
Voice AISpeech AI, Conversational AI
Tools
AI systems for generating, understanding, and manipulating human speech. This includes text-to-speech (TTS)…
Voice CloningVoice Synthesis, Voice Replication
using
Creating a synthetic copy of a specific person's voice from a short audio sample, enabling text-to-speech that sounds…
Voyage AIvoyage-3, domain-specific embeddings
Companies
Embedding model company building specialized vectors for code, legal, finance, and multilingual search. Their models…
VRAMVideo RAM, GPU Memory
Infrastructure
The memory on a GPU, separate from system RAM. AI models must fit in VRAM to run on a GPU. A 7B parameter model in…
Wan-AIWan video models, open-weights video generation
Companies
Alibaba's dedicated video generation initiative, releasing high-quality open-weights video models. Part of Alibaba's…
WatermarkingAI Watermark
safety
Embedding invisible signals in AI-generated content for later detection. Text watermarking subtly biases token…
Weight InitializationXavier Init, Kaiming Init, He Init
training
How neural network weights are set before training begins. Bad initialization can make training fail before it starts…
WeightsModel Weights, Neural Network Weights
Training
The numerical values inside a neural network that get adjusted during training to minimize error. Each connection…
Weights & BiasesW&B, WandB
companies
The dominant MLOps platform for tracking machine learning experiments. W&B lets you log metrics, hyperparameters, model…
WhisperOpenAI Whisper
Models
An open-source speech recognition model released by OpenAI in 2022. Whisper converts spoken audio into text using an…
WindsurfCodeium, Windsurf Editor
companies
An AI-native code editor (formerly Codeium) that competes with Cursor in the AI coding assistant space. Like Cursor…
Word EmbeddingWord2Vec, GloVe, Word Vectors
fundamentals
Dense vector representations of words where words with similar meanings have similar vectors. Word2Vec (2013) and GloVe…
World ModelInternal World Model, Learned Simulator
models
A model that builds an internal representation of how the world works — not just statistical correlations but…
xAIGrok
companies
Elon Musk's AI company, known for Grok models. Has access to X (Twitter) data and one of the largest GPU clusters…
XiaomiMiLM, consumer electronics AI
Companies
One of the world's largest consumer electronics companies, now building its own AI models. MiLM powers features across…
YAMLYAML Ain't Markup Language
Infrastructure
A human-readable data serialization format used extensively in AI and DevOps for configuration files, pipeline…
Zero-shot / Few-shotIn-context Learning
Using AI
Zero-shot means asking a model to do a task with no examples — just the instruction. Few-shot means providing a…
Zhipu AIGLM, ChatGLM, CogView, CogVideo
Companies
Chinese AI company spun out of Tsinghua University. Behind the GLM model family and one of China's leading AI…
ESC
↑↓ Navigate ↵ Open wiki: news: model: Filter