Label
The ground‑truth category or value assigned to a data point used for supervised learning.
Common AI terms beginning with L, defined for advertising professionals.
Find the Flux+Form AI glossary & dictionary to help you make sense of common AI terms. Below you can find an AI Glossary & Dictionary for “L”.
The ground‑truth category or value assigned to a data point used for supervised learning.
A technique that replaces hard labels with a mixture of the correct and incorrect labels to improve generalization.
Data that has been annotated with correct outputs, used for supervised learning.
A data‑processing architecture that handles massive quantities of data by using both batch and real‑time processing.
A generative AI model trained on massive amounts of text to generate human‑like language and perform tasks such as summarization or conversation.
The way an AI tool’s output gradually takes on your individual flavor through sustained use, shaped by your prompting voice, your patterns, and the memory it keeps of how you work, so the same prompt produces a different result for you than for anyone else.
A type of AI model designed to reason through problems step by step before producing an answer, trading speed for greater accuracy on complex tasks like strategy, analysis, and multi-step reasoning.
A compressed representation of input data learned by a model, capturing underlying features.
A technique that normalizes inputs across the features of a layer to stabilize learning.
Ranking leads based on their likelihood to convert using behavioural, demographic and firmographic data.
A hyperparameter that controls how much a model’s weights are updated during training.
A validation method where each observation is used once as the test data while the rest form the training data.
A method for modelling the relationship between a dependent variable and one or more independent variables.
Characteristics extracted from text data, such as syntax or sentiment, used in NLP models.
The practice of structuring and publishing content so that large language models cite your brand in their responses, rather than ranking your pages in traditional search results.
A solution that is better than neighbouring solutions but not necessarily the best overall; models aim to avoid getting stuck in local minima.
Identifying new prospects that resemble existing high‑value customers based on behavioural and demographic traits.
The practice of designing and tuning the iterative feedback cycles in agentic AI systems, where a generation step, an evaluation step, and a revision step repeat until output meets a defined quality threshold or a stop condition is reached.
Platforms that allow marketers to design automated workflows using minimal coding, often integrating AI for decision making.
The dictionary defines the words. The Creative Cadence Workshop builds the practice. Eight weeks of hands-on AI training built for ad agencies and in-house creative teams.
Learn about the workshop →