artificial neural network
Computational learning system
An artificial neural network is a computing system loosely modeled on the brain's structure—layers of interconnected nodes that process information by adjusting the strength of connections between them.
Neural networks learn patterns from data rather than following explicit instructions, making them central to modern machine learning and AI: The neural network was trained on millions of images.
Linear binary classifier algorithm
- backpropagationReverse gradient computation method
- deep learningLayered neural network methods
- cyberneticsCommunication and control science
Internal neural network stages
- nodesPoints of connection
- feedforwardControl anticipating disturbances
- recurrentHappening multiple times
- convolutionalUsing sliding kernel operations
perceptron · activation function · weights · bias · dropout · pooling · encoder · decoder
Reverse gradient computation method
- optimizationMaximizing efficiency or performance
- learning rateUpdate magnitude per iteration
- epochDistinct historical period
- overfittingExcess adaptation to samples
underfitting · regularization · reinforcement learning · hyperparameter · validation · convergence · momentum · cross-validation
Identifying regularities in inputs
- classificationSystematic grouping by traits
- regressionReturn to former state
- image recognitionIdentifying contents of photos
- speech recognitionMachines understanding human talk
natural language processing · computer vision · recommendation system · facial recognition · machine translation · autonomous vehicles · medical diagnosis · drug discovery
Nerve signal crossing point
- neuroplasticityBrain's self-rewiring ability
- parallel processingSimultaneous computational operations
- thresholdPoint of beginning
- excitationStimulation of response
perceptron
Linear binary classifier algorithm
- backpropagationReverse gradient computation method
- deep learningLayered neural network methods
- cyberneticsCommunication and control science
hidden layers
Internal neural network stages
- nodesPoints of connection
- feedforwardControl anticipating disturbances
- recurrentHappening multiple times
convolutional · perceptron · activation function · weights · bias · dropout · pooling · encoder · decoder · transformer
backpropagation
Reverse gradient computation method
- optimizationMaximizing efficiency or performance
- learning rateUpdate magnitude per iteration
- epochDistinct historical period
overfitting · underfitting · regularization · reinforcement learning · hyperparameter · validation · convergence · momentum · cross-validation · fine-tuning
pattern recognition
Identifying regularities in inputs
- classificationSystematic grouping by traits
- regressionReturn to former state
- image recognitionIdentifying contents of photos
speech recognition · natural language processing · computer vision · recommendation system · facial recognition · machine translation · autonomous vehicles · medical diagnosis · drug discovery · generative AI
synapse
Nerve signal crossing point
- neuroplasticityBrain's self-rewiring ability
- parallel processingSimultaneous computational operations
- thresholdPoint of beginning
excitation · inhibition · self-organization · emergent behavior
artificial neural networkhuman brain
Neural organ enabling cognition
brain · cerebral cortex · nervous system · wetware · human mind
artificial neural networklinear regression
Statistical straight-line fitting method
artificial, artificially, artifice, neural, neuron, neuronal, neurons, neurology, neurological