a glimpse into the databases used by artificial intelligence.
Database fot Human training is a collection of works I made between 2017 and 2022 dealing with the databases with which the artificial intelligences were trained. Animated Gifs, Books, Video, and more.
BASIC EMOTION (2017 – 2019)
Basic Emotion presents a series of animated gifs that bring to light the faces of people contained in the databases with which the artificial intelligences were trained. These images were made to be seen by the machine, not by humans, and the project proposes an inversion of the point of view, positioning the viewer on the side of the machine during his training.
At a time when artificial intelligence is becoming ubiquitous and are increasingly present in our everyday life, it is important to ask what materials are basing their training because this process can be a sort of imprinting.
Text: Interview by Bertram Niessen (2019) – Database for Human Training ci riporta alle origini della ricerca sul riconoscimento facciale”
other database:
COLORING BOOK for kids and young AI
Play with the images used by the most advanced computer vision systems!
The images contained in this coloring book are used as a test (Ground truth) to verify the accuracy of the edge detection algorithms used by computer vision systems. They were hand-drawn by thirty anonymous human subjects that traced the main contours of the photographs. These images are part of The Berkeley Segmentation Dataset, a large dataset of natural images that have been manually segmented and provide an empirical basis for research on image segmentation and boundary detection. Many artificial intelligences have been trained on these images for recognizing the objects and main areas of an image, just as children learn to recognize shapes and give it meaning by coloring the white areas of coloring books. This book allows you to use the same materials for training both children and young artificial intelligence.
In this image there is a cat (2021)
video 3:32 – H264 1920×1080, 2021
“In this image there is a cat”, is made with the voices of networkers who were asked to describe in words an image by moving the mouse over the areas they were describing, to gather information for AI training. The choice to remove the original image from the video forces us to focus our attention on voices, ambient noises, mouse tracks, imagining the networkers themselves and their working conditions beyond the images described.
source: https://google.github.io/localized-narratives/
a young woman with bruised face and bloody now holding a stuffed animal mouse in her arms. a woman with a badly bruised face holding a stuffed animal. a woman is hugging a stuffed mouse close to herself a beat up girl holding a teddy bear. an injured woman holding a teddy bear close to her chest
a white cat sitting on a floor under an umbrella. a white cat covering itself with an umbrella a white cat sitting under a white umbrella the white cat is sitting underneath an umbrella a small white cat is sitting under an umbrella.
a cat wearing a yellow bow tie with it’s eyes closed. a cat with a yellow bow tie on. a very fat cat looks into the camera lense a black cat wearing a yellow bow tie. a black cat wearing a yellow bow tie.
a cat that is looking out the window at a turkey. a cat is watching a turkey through a window. an orange cat sitting on top of a window sill next to a bird. a cat is looking out the window at a bird in the yard. cat is watching a bird through a window.
a cat sits and watches tv, which has a cat on it.
a brown cat watching the tv about a dog
a cat sitting in a living room looking at a television
a dog sitting and watching a dog on television.
the cat is fixated as it watches the cat television show.