Microsoft Azure AI Fundamentals (AI-900) Practice Exam

When is a model considered to be overfitting?

1. When it performs well on training data but poorly on validation data

2. When it has high accuracy on both training and validation data

3. When it uses less data than necessary

4. When it cannot learn any patterns from the data

The correct answer is: When it performs well on training data but poorly on validation data

A model is considered to be overfitting when it performs exceptionally well on the training data but shows poor performance on validation data. This indicates that the model has learned the noise and details in the training data to the extent that it fails to generalize to new, unseen data. In essence, the model has memorized the training examples instead of learning the underlying patterns that would allow it to make accurate predictions on different datasets. Overfitting typically occurs when the model is too complex relative to the amount of training data available. This results in a model that captures noise and outliers as if they were valid patterns, leading to high accuracy on known data while failing to perform well on new data. Hence, a key indicator of overfitting is the disparity between training and validation performance, where the training accuracy is high, but validation accuracy is significantly lower. This is a critical concept in machine learning, as it underscores the importance of balancing model complexity with adequate training data to ensure the model can generalize effectively.