Neural Network
Subject: Computer and Technology noun /ˈnʊr.əl ˈnɛtˌwɜːrk/ 1 definition
- artificial-intelligence
- machine-learning
- deep-learning
- computer-science
-
Definition (noun)
Entry #207 · UpdatedA Neural Network (NN) is a computational model inspired by biological brains that transforms input data through layers of interconnected nodes—called artificial neurons—to learn patterns and produce predictions or classifications. Each connection carries a tunable weight, and each node applies an activation function to its weighted inputs; during training, optimization methods such as backpropagation and gradient descent iteratively adjust these weights to minimize output error. Unlike traditional rule-based programs that execute explicitly coded logic, a neural network infers statistical representations directly from training data, enabling it to generalize across complex, unstructured domains like vision, audio, and natural language.
Common Types
Feedforward Neural Network (FNN): Routes signals strictly forward from input through hidden layers to output without feedback loops, making it well-suited for tabular classification and regression.
Convolutional Neural Network (CNN): Applies learnable spatial filters across grid-structured inputs to extract hierarchical features, excelling in image recognition and computer vision.
Recurrent Neural Network (RNN): Retains internal state across sequential steps via feedback connections—such as in Long Short-Term Memory (LSTM) variants—to model time-series data, speech, and text.
Transformer: Uses self-attention mechanisms rather than recurrence to weigh relationships across entire sequences in parallel, powering modern large language models and multimodal systems.
Example
A hospital deploys a convolutional neural network trained on 200,000 labeled chest radiographs to screen incoming X-ray scans for early signs of pneumonia. As each scan passes through the network's layers, learned filters highlight subtle lung opacities and output a calibrated probability score that flags high-risk cases for priority radiologist review.