Neural network

📖 Full definition

A neural network consists of layers of units ("neurons") connected by adjustable weights. Training means modifying those weights so the network produces the right outputs from given inputs. It is the foundation of almost every recent AI: vision, language, audio, prediction.

💡 Concrete example

A facial recognition neural network typically has 50 to 200 million adjustable parameters, trained on tens of millions of faces.

🔄 Synonyms

NN · ANN

🔗 See also

• Deep learning →
• CNN (Convolutional Neural Network) →
• Facial embedding →

📰 Articles mentioning this

• Who Does My Child Look Like? What Family Resemblance Actually Reveals →
• Adoption: Why Do You Sometimes "Look Like" Your Adoptive Parents? →
• 15 Celebrities Who Look Strikingly Like Their Parents →
• How Does Our Facial Recognition AI Work? →