Advanced · Neural Networks
Neural Turing machine (NTM)
Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.
TL;DR. A neural network combined with external memory, capable of learning algorithms.
Technical Definition
A neural network combined with external memory, capable of learning algorithms.
How it works
A Neural Turing Machine (NTM) is a type of recurrent neural network architecture that augments a standard neural network controller with an external memory component. It can learn to interact with this memory using differentiable attention mechanisms, enabling it to perform tasks that require algorithmic capabilities, such as copying and sorting. NTMs represent a step towards neuro-symbolic integration.
Related Concepts
- Gradient Descent — An optimization algorithm that iteratively adjusts model parameters by moving in the direction of steepest decrease of the loss function.
- Recurrent Neural Network (RNN) — A neural network with loops that maintain hidden state, designed to process sequential data like text and time series.