80 Million Tiny Images is a dataset intended for training machine learning systems constructed by Antonio Torralba, Rob Fergus, and William T. Freeman in a collaboration between MIT and New York University. It contains 79,302,017 32×32 pixel color images, scaled down from images extracted from the World Wide Web in 2008 using automated web search queries on a set of 75,062 non-abstract nouns derived from WordNet. The words in the search terms were then used as labels for the images.[1] The researchers used seven web search resources for this purpose: Altavista, Ask.com, Flickr, Cydral, Google, Picsearch and Webshots.[1]
The dataset was motivated by non-parametric models of neural activations in the visual cortex upon seeing images.[2]
The CIFAR-10 dataset uses a subset of the images in this dataset, but with independently generated labels, as the original labels were not reliable. The CIFAR-10 set has 6000 examples of each of 10 classes and the CIFAR-100 set has 600 examples of each of 100 non-overlapping classes.[3]
The 80 Million Tiny Images dataset was retired from use by its creators in 2020,[4] after a paper by researchers Abeba Birhane and Vinay Prabhu found that some of the labeling of several publicly available image datasets, including 80 Million Tiny Images, contained racist and misogynistic slurs which were causing models trained on them to exhibit racial and sexual bias.[5][6] Birhane and Prabhu also found that the dataset contained a number of offensive images.[6][7] Following the release of the paper, the dataset's creators removed the dataset from distribution, and requested that other researchers not use it for further research and to delete their copies of the dataset.[4]