# What is Generalization? Definition & Technical Mechanics

> **Source:** [SPIDITS AI AI Glossary](https://spidits.com/ai-glossary)  
> **Category:** Model Training  
> **Canonical Citation:** [https://spidits.com/ai-glossary/generalization](https://spidits.com/ai-glossary/generalization)

## Definition
Generalization is a machine learning model's ability to make accurate predictions on new, unseen test data that was not present in the dataset used to train the network.

## Architectural Mechanics & Deep Dive
Generalization is the ability of a machine learning model to make accurate predictions on new, unseen data that was not part of its training set. A model generalizes well when it learns the underlying patterns rather than memorizing the training noise (overfitting). Evaluating generalization requires measuring model performance on independent test datasets.

## Industry Applications & Real-World Use Cases
Validation curve testing, model capability audits, and real-world deployment checks.

## Frequently Asked Questions

### How do you measure a model's generalization capability?
By checking its accuracy metrics on a completely held-out test dataset that the model weights never saw during optimization.

### What prevents a model from generalizing well?
Overfitting training noise (memorizing details rather than learning general concepts) or a significant shift in data distributions (out-of-distribution inputs).

## Related Terminology
- [overfitting](https://spidits.com/ai-glossary/overfitting)
- [underfitting](https://spidits.com/ai-glossary/underfitting)
- [validation-data](https://spidits.com/ai-glossary/validation-data)
- [out-of-distribution](https://spidits.com/ai-glossary/out-of-distribution)

---
*From the SPIDITS AI Knowledge Base (336+ terms indexed). Explore full technical definitions: [https://spidits.com/ai-glossary](https://spidits.com/ai-glossary)*