inference economics
AI inference economics
AI inference economics examines the revenue, cost, utilization, pricing, and margins involved in serving trained models.
AI Terminology
Search clear AI terminology definitions, acronyms, related concepts and the reviewed videos that explain them.
Showing 461–480 of 705 terms
Clear filtersinference latency
AI inference latency is the time between submitting model input and receiving the required output or response milestone.
inference pod
AI inference podAn AI inference pod is a deployable group of compute and software resources that serves requests for an AI model.
infrastructure investment
AI infrastructure investmentAI infrastructure investment funds the chips, data centers, power, networking, storage, and operations needed to develop and run AI systems.
innovation acceleration
AI innovation accelerationAI-accelerated innovationAI innovation acceleration is the increase in the speed or rate of creating and applying new knowledge, products and processes with AI.
intelligent document processing
AI document processingIDPIntelligent document processing uses AI to classify, extract and validate information from documents for downstream workflows.
intent-based artificial intelligence workflow
An intent-based AI workflow starts from a desired outcome and determines the coordinated tasks needed to reach it.
inter-token latency
ITLInter-token latency is the time between consecutive output tokens while an AI model is generating a response.
internet-exposed system
An internet-exposed system is a device, application or service that can be reached directly or indirectly from the public internet.
job exposure
AI job exposureAI job exposure estimates how much of an occupation's tasks could be affected by current or emerging AI capabilities.
job task bundle
A job task bundle is the connected set of activities, decisions and responsibilities that together make up an occupation or role.
key-value cache-aware routing
KV-cache-aware routingKey-value cache-aware routing sends a model request to a worker that can reuse relevant cached attention state while also considering current load.
knowledge distillation
KDKnowledge distillation trains a smaller or different model to reproduce useful behavior learned from the outputs or internal signals of a more capable teacher model.
knowledge graph
A knowledge graph represents entities, concepts and facts as connected nodes and relationships that software can query and traverse.
language model logit
LLM logitA language model logit is an unnormalized score assigned to a possible next token before those scores are converted into probabilities.
language model sampling distribution
LLM sampling distributionA language model sampling distribution assigns a probability to each possible next token after model scores and decoding controls are applied.
language model token entropy
LLM token entropyLanguage model token entropy measures uncertainty in the probability distribution over possible next tokens.
language model watermark paraphrase attack
LLM watermark paraphrase attackA language model watermark paraphrase attack rewrites watermarked text to preserve meaning while weakening the original keyed token pattern.
language model watermarking
LLM watermarkingLanguage model watermarking modifies token selection so generated text contains a hidden statistical pattern that an authorized detector can test for later.
language-mediated multi-robot coordination
Language-mediated multi-robot coordination uses structured or natural-language messages to organize responsibilities and handoffs among robots.
Knowledge graph
Explore AI concepts
Follow general terms into specialised sub-terms. Select any node to open its definition.