Purpose
To provide a method of distinguishing data origins. Many artists would like to distinguish between model and instrument generated art. Many machine learning models also want to ensure non-model data is input to ensure variety in training data. These labels will enable filtering based on the type of data one would like to see while also providing a standardized disclaimer about the data shown, ensuring information is interpreted based on origin.
Implementation
For use in physical media, these tags can be printed and pasted onto data or art. Stickers of these tags can also be purchased on this site.
For use in digital media, these tags can be downloaded and placed over data or art. These tags can also be placed at the top of articles to indicate the type of data used. Finally, for NFTs these tags can be added as a form of metadata such as through groups on OpenSea.
Tags
Instrument generation
Data, information, content, or other phenomena recorded by a device, originating from the natural world. This includes paintings, recordings of analog musical instruments, drawings with pen and paper, film photography, and most scientific sensing data from instruments such as thermocouples, pressure transducers, and digital cameras.
Model generation
Predicted data based on a model. This includes images, text, values and any other data that are produced by a model based on instrument data. Computation produces the data, not merely processes it. Information covered in this category includes learning model generated data as well as model predictions for scientific findings.