📖 AI Glossary
Look up AI terms, companies, and services used across AI Note Lab.
-
Term
Quantization
Quantization is a representative lightweight technique that reduces the size and amount of calculations by lowering the numerical precision of the model.
-
Term
RAG
retrieval augmented generation
RAG is a technique that retrieves relevant documents before generating an answer, helping reduce hallucinations.
-
Term
Red Team
red teaming
Red Team is a verification activity that detects risks in advance by deliberately attacking AI’s loopholes.
-
Term
Reinforcement learning
RL
Reinforcement learning is a learning method that allows you to learn through trial and error by giving rewards if you do well and penalties if you do not. This is the core technique of AlphaGo and its inference model.
-
Term
RLHF
Human Feedback Reinforcement Learning
RLHF is a training technique in which humans score AI answers and refine them to provide more useful and safer answers. It is considered a key technology that made ChatGPT usable.
-
Term
RPA
robotic process automation
RPA is a technology in which software robots replace repetitive computer tasks (input, copy, transfer) previously performed by humans.
-
Term
Scaling law
The rule of thumb is that as the amount of model, data, and calculation increases, performance predictably improves. This is the theoretical basis for AI investment competition.
-
Term
Singularity
This refers to the hypothetical point when AI surpasses human intelligence and humans cannot control the speed of its development.
-
Term
Sort
alignment
Sort is a field of research that makes AI behave in line with human intentions and values. This is a core concept in AI safety discussions.
-
Term
Sovereign AI
national AI
This refers to national-level AI created with domestic data and language rather than relying on foreign technology. These are Korea’s domestic AI policy keywords.
-
Term
Stargate
Stargate is a large-scale AI data center construction project promoted by OpenAI, Softbank, and Oracle. It symbolizes the AI infrastructure competition worth hundreds of billions of dollars.
-
Term
STT
speech recognition, speech to text
STT is the art of transcribing words into writing. It is used to automatically create meeting minutes and create subtitles.
-
Term
Supervised learning
Supervised learning is a learning method that teaches AI using data with correct answers. This is done by labeling the photo “cat” and training it.
-
Term
Synthetic data
The data generated by AI is used again for AI learning. It is an alternative to the depletion of real data and a subject of debate over quality degradation.
-
Term
System prompt
System prompt refers to basic instructions given to AI before starting a conversation. AI’s tone and rules are set here.
-
Term
Token
A token is the smallest text unit an AI model processes, and it is commonly used to calculate usage limits and costs.
-
Term
TPU
tensor processing unit
TPU is an AI-specific chip designed by Google itself. It is attracting attention as a rival to NVIDIA GPUs.
-
Term
Transformer
Transformer is an AI structure announced by Google in 2017, and is the foundation technology for most modern generative AI, including ChatGPT.
-
Term
TTS
speech synthesis, text to speech
TTS is a technology that reads text in a natural voice. It is the foundation of AI dubbing and audiobooks.
-
Term
Turing Test
If you cannot tell whether the person you are talking to is an AI or a human, it is a classic criterion to assume that you are intelligent.
226 entries · page 11 of 12