Green BOA Analyzes ML Compression's Environmental Break-Even Point

This research, titled "Green BOA," summarizes findings from two summer internship projects at the University of Manchester. The core focus is on determining the environmental break-even point for machine learning-based data compression algorithms. The study uses a specific ML-based lossless compression algorithm as an example. It meticulously compares the estimated carbon-equivalent emissions associated with the infrastructure required for both ML training and subsequent inference processes. These carbon costs are then weighed against the carbon-equivalent savings achieved through reduced disk storage requirements due to the compression. The ultimate goal is to identify the point at which the environmental benefits of reduced storage outweigh the energy consumption and emissions from developing and running the ML compression model.