Machine Learning Beginner
training-serving skew
The difference between a model's performance during training and that same model's performance during serving.
Plain English Explanation
The difference between a model's performance during training and that same model's performance during serving.
How is it used?
Practitioners refer to training-serving skew when building, training, or evaluating machine learning systems. It appears in research papers, product documentation, and technical discussions about AI capabilities and limitations.