# Definition
A neural network is composed of the interconnection of fundamental units called **neurons**.
Every **neuron** uses the input values received and process this values to generate an output. Then, every neuron has the following inputs represented by:
$
X = \{x_1, x_2, \dots, x_n\}
$
Every input is pondered by the weight values defined by:
$
W^i = \{w_1^i, w_2^i, \dots, w_n^i\}
$
We should interpret the value $w_j^i$ as the weight or importance of the input value $x_j$ that reaches neuron $i$ from neuron $j$.
We use an [[Input or combination function]]. And we apply an [[Activation function]] to generate a new output named $y_i$.
![[Pasted image 20240902104838.png]]
---
## References
- [[Deep learning - Anna Bosch Rué Jordi Casas Roma Toni Lozano Bagén]]
- [[DeepLearningPrincipios.pdf]]