Neural Network
A neural network is a function built from layers of simple units. Each unit multiplies its inputs by weights, adds them, passes the sum through a small nonlinearity, and forwards the result. Training adjusts the weights until the whole stack produces useful outputs.
No jargon, no math
In plain words, what is it?
Picture a panel of judges scoring a dive: each judge weighs different things and shouts a score, and the next panel weighs those shouts. A neuron does exactly that with numbers: it takes inputs, multiplies each by a weight (how much it trusts it), adds a bias, and pushes the sum through an activation that decides how loudly it passes the signal. The input layer receives raw values, the hidden layers build patterns layer after layer, the output layer gives the answer. Nothing in there understands anything; it is arithmetic, billions of multiplications. Learning means nudging every weight in the direction that reduces the error, millions of times. The boundary: single weights mean nothing readable, so when the network is wrong you often cannot ask why.
- 01
Input. Raw values enter: pixels, numbers, fragments of words.
- 02
Weighted sum. Each neuron multiplies its inputs by weights and adds a bias.
- 03
Activation. A small nonlinearity decides how strongly the signal passes on.
- 04
Next layer. Outputs feed the next layer; patterns get more abstract each time.
- 05
Output and correction. The last layer answers; the error nudges every weight.
Terms that live nearby.
Deep Learning
Deep learning is machine learning with neural networks stacked in many layers, so the system learns the features by itself instead of a human hand-picking them. It is the technique behind today's language, image and speech models.
ReadParameter
A parameter is one of the internal numbers a trained model stores: billions of them, each tuned during training, and together they encode everything the model learned.
ReadTraining
Training is the process that turns an untrained model into a working AI: you feed it thousands or millions of examples, compare its guesses with the correct answers, and nudge its internal settings until the guesses get good.
Read§ 04
Want this running in your stack?
We design and ship AI systems with permissions, logs and human approval. It starts with a 30-minute scoping call.