AI Terms Dictionary

Definition

Randomness is fundamental in AI for initializing model weights, shuffling datasets, and introducing stochasticity during training to prevent overfitting. Since computers are deterministic, AI systems use pseudo-random number generators (PRNGs) seeded with specific values to produce sequences that appear random. Controlling this randomness via seeds ensures reproducibility of experiments and model results.

Summary

The property of lacking a predictable pattern, often simulated in AI through pseudo-random number generation algorithms.

Key Concepts

• Seed Value
• Stochasticity
• Pseudo-Random
• Reproducibility

Use Cases

• Weight initialization in neural networks
• Dropout regularization
• Exploration in reinforcement learning

Code Example

1
2
3

import numpy as np
np.random.seed(42)
print(np.random.rand())

Related Terms

• Noise
• Entropy
• Distribution
• Seed