← Glossary
LLM & Generative AI
Large language models, generative AI, agents, RLHF, multimodal
44 terms
agentic access management
행위적 접근 관리
Agentic Access Management (AAM) is an emerging discipline focused on managing how AI agents interact with enterprise sys…
Agentic workflows
에이전트 워크플로우
Agentic workflows are dynamic workflows in which multiple specialized AI agents collaborate to plan, reason, use tools, …
AI robotics
AI 로보틱스
AI robotics is the integration of artificial intelligence (AI) technologies into robotics, enabling robots to perceive, …
AI-native framework
AI 네이티브 프레임워크
An AI-native framework refers to systems or processes designed from the ground up with AI as a core component, with AI c…
Anthropic
앤트로픽
Anthropic is an American AI safety and research company based in San Francisco. It is best known for building Claude, a …
AUC (Area Under the Curve)
곡선 아래 면적
AUC represents the area under the ROC curve and is used as a metric to evaluate the performance of a classification mode…
Bedrock
베드록
Amazon Bedrock is an AWS fully managed service that provides access to high-performing foundation models from multiple p…
BERT
버트
BERT (Bidirectional Encoder Representations from Transformers) is a transformer-based language model introduced by Googl…
edge deployment
에지 배포
Edge deployment means running AI models or apps close to where data is created — for example on factory lines, inside re…
embedding
임베딩
An embedding is a mathematical method of representing complex data in a lower-dimensional space to make it easily proces…
foundation model
기초 모델
A foundation model is a large-scale AI model pre-trained on vast amounts of data and can be adapted for various tasks.
Gemini
제미나이
Gemini is Google’s family of multimodal generative AI models and the chatbot/app built on them. Unlike text-only systems…
Gemma 4
젬마 4
Gemma 4 is the latest generation of Google DeepMind’s lightweight, open-weight AI models designed to run efficiently on …
GPT-4o
GPT-4o
GPT-4o is OpenAI’s latest large language model that can handle text, speech, and images all at once. It’s designed to be…
grouped-query attention
그룹 쿼리 어텐션
Grouped-query attention is a method used in large language models (LLMs) and transformer-based AI systems to process sev…
hallucination
환각
AI hallucination is a phenomenon where a large language model or other generative AI system produces outputs that seem c…
inference cost
추론 비용
Inference cost is the ongoing cost of running data through a trained AI model to produce an output—like a prediction, ge…
inference latency
추론 지연 시간
Inference latency is the actual time it takes for an AI model to process an input and return an output. It typically ref…
Latent MoE
잠재적 전문가 혼합
Latent MoE (Latent Mixture of Experts) is a variant of sparse Mixture of Experts where each expert operates in a smaller…
LLM
대규모 언어 모델
A Large Language Model (LLM) is a deep learning model trained on massive text corpora to understand and generate human l…
mixture of experts
전문가 혼합
A mixture of experts is an AI architecture that combines several specialized models (called 'experts') and decides which…
Model parallelism
모델 병렬 처리
Model parallelism is a technique where a large AI model is partitioned across multiple devices (such as GPUs), with each…
multi-agent system
다중 에이전트 시스템
A multi-agent system is a network of multiple artificial intelligence agents that interact within a shared environment, …
multi-hop retrieval
다중 홉 검색
Multi-hop retrieval is a technique where an AI system answers complex queries by sequentially retrieving and connecting …
multi-stage training
다단계 학습
Multi-stage training is a method for developing AI models—especially large language models (LLMs)—by progressively impro…
multimodal model
멀티모달 모델
A multimodal model is an artificial intelligence model capable of simultaneously understanding and processing multiple t…
NLP
자연어 처리
Natural Language Processing (NLP) is a branch of artificial intelligence that enables computers to read, understand, and…
Nvidia
엔비디아
Nvidia is a technology company best known for its graphics processing units (GPUs) and a full-stack AI platform that inc…
on-device AI
온디바이스 AI
On-device AI means running artificial intelligence directly on your own device—like a phone, laptop, or tablet—instead o…
open-source LLM
오픈소스 대형 언어 모델
An open-source large language model (open-source LLM) is a type of AI language model whose underlying code and trained d…
OpenAI
오픈에이아이
OpenAI is an artificial intelligence research organization and platform founded in 2015 that aims to develop safe and be…
output tokens
출력 토큰
Output tokens are pieces of text generated by an AI model in response to input, where the model predicts the next most l…
post-training
후 훈련
Post-training refers to the set of processes and techniques applied to a machine learning model after it has been initia…
pre-training
사전 훈련
Pre-training is the process of initializing a machine learning model by training it on a large, generic dataset before f…
PyTorch
파이토치
PyTorch is an open-source deep learning framework used to build and train neural networks. With its Python-based intuiti…
RAG
검색 증강 생성
Retrieval-Augmented Generation (RAG) is an architecture that improves LLM outputs by retrieving relevant information fro…
real-time inference
실시간 추론
Real-time inference refers to the process where a trained machine learning model accepts live input data and generates p…
recurrent mechanism
순환 메커니즘
A recurrent mechanism refers to an architectural design in AI models where the output from a previous step is fed back a…
RLHF
인간 피드백 기반 강화학습
Reinforcement Learning from Human Feedback is a method where AI learns better behaviors by using human-provided evaluati…
Self-Attention
셀프 어텐션
Self-attention is a mechanism where each element in an input sequence compares itself with all other elements to compute…
supervised fine-tuning
지도 미세 조정
Supervised fine-tuning is the process of further training a pre-trained AI model using additional labeled data, where hu…
TensorFlow
텐서플로우
TensorFlow is an open-source machine learning and deep learning framework developed by the Google Brain team, designed f…
Transformer
트랜스포머
A Transformer is a neural network architecture that uses self-attention so each token in a sequence can look at every ot…
vision-language model
비전-언어 모델
A vision-language model is an artificial intelligence model designed to simultaneously understand and process both visua…