There's general consensus in the Machine Learning community that models can and have made biased decisions against traditionally marginalized groups. Ethical AI researchers from Dr. Cathy O'Neil to Dr. Joy Buolamwini have gone to great lengths to establish a pattern of faulty decision making rooted in biased and unrepresentative data that result in serious harms. Unfortunately, our "intelligent" learning algorithms are only as smart, capable and ethical as we make them and we are only at the beginning of understanding the long term effects of biased models. Fortunately, there are many strategies already at our disposal that we can use to mitigate harms when they arise. Today, we will focus on a very powerful strategy: Model Rollbacks through Versioning.

A new version of pins is available on CRAN today, which adds support for versioning your datasets and DigitalOcean Spaces boards! As a quick recap, the pins package allows you to cache, discover and share resources. You can use pins in a wide range of situations, from downloading a dataset from a URL to creating complex automation workflows (learn more at pins.rstudio.com). You can also use pins in combination with TensorFlow and Keras; for instance, use cloudml to train models in cloud GPUs, but rather than manually copying files into the GPU instance, you can store them as pins directly from R. You can find a detailed list of improvements in the pins NEWS file. To illustrate the new versioning functionality, let's start by downloading and caching a remote dataset with pins.