Occam's Razor (in ML)

This principle suggests that, when faced with multiple models that explain the data equally well, the simplest model is usually the best. It encourages choosing models with fewer assumptions or parameters to avoid unnecessary complexity and potential overfitting.