Training performance continues improving while validation performance stagnates or worsens. Regularization, more representative data, simpler models, and earlier stopping can reduce the gap.
Overfitting occurs when a model learns training-specific patterns that do not generalize well to new data.
Training performance continues improving while validation performance stagnates or worsens. Regularization, more representative data, simpler models, and earlier stopping can reduce the gap.