-
Perceptron: The Perceptron is a single-layer neural network model that can perform binary classification. It consists of inputs, weights associated with those inputs, a weighted sum function, and an activation function. Although basic, it is the foundation on which more complex neural network models are built.
-
Single Hidden Layer Neural Network (Feedforward Neural Network): This neural network consists of an input layer, a hidden layer and an output layer. You can learn to train this network for more complex problems that cannot be handled effectively by a perceptron.
-
Convolutional Neural Network (CNN): Especially useful for image processing, a CNN uses convolutional layers to detect spatial patterns in the input data.
-
Recurrent Neural Network (RNN): This type of network is ideal for sequential or time series data. You can learn to predict or generate sequences of data.
-
Autoencoder: An autoencoder is a neural network used to learn efficient representations of data. It is useful for dimensionality reduction and generating data similar to the input.
-
Generative Adversarial Network (GAN): GANs are used to generate new data that is similar to a training data set. You can learn to generate realistic images, for example.
-
Long Short-Term Memory (LSTM) Recurrent Neural Network: A type of recurrent network improved to handle problems of long-term dependencies in sequential data.
-
Pretrained Neural Network (Transfer Learning): You can learn to use pretrained neural networks for specific tasks, such as image recognition. This is especially useful when you have access to larger data sets.
jonas1ara / fsharp-nns-playground Goto Github PK
View Code? Open in Web Editor NEWA collection of neural networks for the three people interested in AI using F#
License: MIT License