Tag model evaluation

MIT researchers reveal critical flaws in machine learning models when applied to new data, emphasizing the need for more nuanced evaluation methods.