Why does a model need data it has not trained on?
Unseen evaluation data helps measure generalization. Without it, a system may appear successful because it memorized examples or exploited accidental patterns in the training set.
Explore This Question
Search for additional explanations using these tools:
More FAQs in How did machine learning contribute to the evolution of AI?
No. Many applications combine learned models with rules, search, constraints, or planning. Different methods solve different parts of a system's overall task.
Data leakage occurs when training or evaluation uses information that would not be available in the intended prediction setting, making reported performance misleading.
Still have questions?
Browse all our FAQs or reach out to our support team