# What is Inductive Bias? Definition & Technical Mechanics

> **Source:** [SPIDITS AI AI Glossary](https://spidits.com/ai-glossary)  
> **Category:** Mathematical Foundations  
> **Canonical Citation:** [https://spidits.com/ai-glossary/inductive-bias](https://spidits.com/ai-glossary/inductive-bias)

## Definition
Inductive Bias refers to the assumptions a machine learning algorithm uses to predict outputs for unseen inputs. It prioritizes specific solutions based on the structural design of the model.

## Architectural Mechanics & Deep Dive
Inductive bias refers to the set of assumptions a machine learning algorithm uses to predict outputs for unseen inputs. It represents the model's structural preference for certain solutions. For example, CNNs assume that local pixels are highly related (spatial locality), while RNNs assume temporal coherence across sequences. Inductive bias is critical for enabling models to generalize from limited training examples.

## Industry Applications & Real-World Use Cases
Neural network design, algorithm comparison, and generalization audits.

## Frequently Asked Questions

### What is the inductive bias of a CNN?
Spatial locality (pixels close together are related) and translation invariance (objects are the same regardless of where they are in the frame).

### Why do Transformers have less inductive bias than CNNs?
Because their attention mechanism models relationships between all tokens globally, requiring more training data to learn spatial arrangements.

## Related Terminology
- [cnn](https://spidits.com/ai-glossary/cnn)
- [transformer](https://spidits.com/ai-glossary/transformer)
- [vision-transformer](https://spidits.com/ai-glossary/vision-transformer)

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