Backpropagation
The training algorithm for neural networks that computes gradients of the loss function and updates weights accordingly.
Common AI terms beginning with B, 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 “B”.
The training algorithm for neural networks that computes gradients of the loss function and updates weights accordingly.
A method for representing text by the frequency of each word, ignoring grammar and word order.
An ensemble technique that trains multiple models on resampled subsets of data and averages their predictions.
The capacity of a system to process data; in marketing this affects how quickly models can handle large campaign datasets.
A statistical method that updates the probability of a hypothesis as more evidence becomes available.
A probabilistic graphical model representing conditional dependencies among variables.
An optimization technique that uses Bayesian inference to find the optimal parameters for a function or model.
Analysis of customer behaviour data to uncover patterns that inform marketing strategies and personalization.
Measuring model performance against standard data sets or industry benchmarks to ensure quality and fairness.
Systematic errors that produce unfair outcomes due to skewed data or model design; marketers must mitigate bias to ensure equitable campaigns.
The balance between model complexity (variance) and generalization (bias) that determines predictive accuracy.
Extremely large data sets that require specialized tools for storage, analysis and processing.
A predictive task where the output is one of two classes, such as churn vs. not churn.
The number of bits used to store each pixel, affecting the range of colours or shades in images.
Models whose internal workings are difficult to interpret; this raises concerns about transparency and accountability.
Software that automates tasks by interacting with systems or users, including chatbots and web crawlers.
Identifying networks of compromised devices used to commit fraud or spread malware.
Identifying the edges of objects within images to assist with segmentation and visual understanding.
Technology that enables direct communication between the brain and external devices; relevant to future immersive marketing experiences.
The use of AI to estimate how marketing exposure affects brand metrics like awareness, favourability and intent.
Practices and technologies that protect brands from having their ads appear in inappropriate contexts; AI helps classify content and enforce guidelines.
A simple method of problem solving that tries all possible solutions; useful as a baseline but inefficient for large search spaces.
Information collected about signals indicating a prospect’s likelihood to purchase, used to prioritize outreach and personalization.
A tokenization method used in language models that segments text into subword units for efficient learning.
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 →