Exploring Galaxy Morphology with Autoencoders
Using variational autoencoders to explore the latent space of a large-scale galaxy image dataset.
keras
Latent Space Visualisation
Some techniques for visualising the latent space of a convolutional neural network.
Features and Filters: Looking Inside CNNs
Visualising the feature maps and convolutional filters of one of my galaxy classification CNNs
Hyperparameter Tuning
A guide to tuning convolutional neural network architectures with Optuna and KerasTuner
Visualising CNN Feature Maps
A brief guide to coding and training a basic CNN with the MNIST dataset, and plotting the feature maps.