The soluzione AI glossary helps you understand AI terminology.
ChatGPT, Chatbot, RAG and LLM… You have probably heard at least one of these terms in a meeting or during a coffee break with colleagues. With the rapid development of AI technologies, more and more new terms are finding their way into the world of work. The soluzione AI glossary will help you to familiarize yourself with the most important terms. The easy-to–understand explanations will help you to keep up with conversations and gradually expand your vocabulary for everyday working life in the future.
Artificial intelligence (AI):
Computer system that is able to perform tasks that would actually require human intelligence.
Machine Learning (ML):
Learning based on examples, empirical values or large amounts of data. These are analyzed and patterns are identified.
Artificial neural networks (ANN):
A structure of neurons and layers modeled on the human brain that enables information to be processed and linked.
Deep Learning (DL):
AI uses neural networks to learn from itself. Systems analyze, link and come to decisions that humans have not explicitly programmed it to make. Deep learning models are able to recognize and learn very deep correlations.
Large Language Models (LLM):
A LLM has been trained as an artificial neural network using huge quantities of texts.
Natural Language Processing (NLP):
Natural language processing; a branch of AI that deals with the analysis, understanding, and generation of human language by machines.
Supervised learning:
Machine learning, in which a model is trained using data that contains inputs and the corresponding desired outputs.
Unsupervised learning:
Learning approach in which a model recognizes patterns and structures in data without specific target values being specified.
Reinforcement Learning (RL):
Learning through reinforcement; a model learns through interaction with an environment by being rewarded for correct decisions and punished for wrong ones.
Algorithm:
Step-by-step instructions or set of rules used to solve a problem or perform a specific task.
Big Data:
Very large and complex amounts of data that are difficult to process with traditional data processing tools.
Training:
The process by which an AI model is optimized based on training data in order to perform certain tasks better.
Bias (distortion):
Systematic deviation or bias in data or models that may lead to erroneous or inaccurate results.
Cloud computing:
Provision of IT resources and services such as storage space, computing power and applications via the Internet.
Computer Vision:
Sub-area of AI that enables machines to understand and interpret images and videos.
Data preprocessing:
Steps for the preparation of raw data, e.g. Cleansing, transformation or normalization to make them usable for AI models.
Robotic Process Automation (RPA):
Technology that uses software bots to automate repetitive tasks.
Turing test:
Test that measures whether a machine can imitate human behavior so well that a human cannot tell the difference.
Prompt:
Input or request given to an AI system to obtain an answer or solution. A precise prompt usually leads to better results.
Generative AI:
AI that creates new content such as text, images, videos, or music based on the data it has been trained with.
Fine-tuning:
Adaptation of an AI model with specific data to optimize it for a specific task.
Text-to-text models:
Models that process text input and convert it into other texts, e.g. Summaries, translations or answers to questions.
Text-to-image models:
Models that can generate images from text input, e.g. for visualizing concepts or designs.
Automated texts:
AI-generated texts such as product descriptions, emails, or reports that can increase efficiency and save time.
Text summary:
Function of AI that shortens long texts to the most important points. Especially useful for reports or articles.
Semantic search:
Searches based not just on keywords, but on the meaning of words to provide more relevant results.
Chatbot:
AI system that responds to user input in natural language and is often used in customer service or information services.
Automated translation:
AI-generated translations that deliver fast and often high-quality results.
Image classification:
AI function that analyzes images and assigns categories to them, e.g. “dog”, “car” or “landscape”.
Sentiment analysis:
AI tool that analyzes texts to identify the prevailing mood (e.g. positive, negative, neutral).
Automation:
Use of AI to make repetitive or time-consuming tasks more efficient, such as email management or data analysis.
Content creation:
AI tools that generate content such as blog posts, social media posts or product descriptions.
Language generation:
Function that makes it possible to create human-sounding texts or spoken content, e.g. for podcasts or presentations.