NAIROBI, Kenya, June 17 – Artificial intelligence (AI) continues to gain traction, and there is a critical aspect that comes with it: breaking down the terminologies associated with AI.
Since generative AI became mainstream in late 2022, understanding its evolving terminology is crucial. Tech giant Microsoft breaks it down, and here are ten key AI terms to keep you updated:
Reasoning and Planning: AI now mimics human reasoning and can devise plans to solve complex problems.
Training and Inference: Training is the AI’s learning phase from data, while inference is applying that learned knowledge to make predictions on new data.
Small Language Models (SLM): Compact versions of large language models that work efficiently on devices without internet connections.
Grounding: The process of anchoring AI models with real-world data to improve accuracy and reduce errors.
Retrieval Augmented Generation (RAG): Enhances AI accuracy by integrating external data sources without retraining the model.
Orchestration: Manages the sequence of AI tasks to ensure coherent and accurate responses, maintaining context across interactions.
Memory: AI uses orchestrated instructions to temporarily store and utilize context from previous interactions, aiding coherent responses.
Transformer and Diffusion Models: Transformers are key for language generation, while diffusion models gradually refine images, showcasing different AI applications.
Frontier Models: Cutting-edge AI systems with advanced capabilities often surprise developers with their potential.
Graphics Processing Unit (GPU): Essential for AI due to their ability to perform parallel processing, powering the most advanced models.
