Domain-adaptation-deep-learning

Task	Description	Link
Rotation Domain Adaptation	Testing domain adaptation using ResNet-50 on the CIFAR dataset, where the source distribution will be the original images and the target distribution will be the transformed version of the dataset created by rotating images at certain angles. Comparing the results with L2 regularized, L1 regularized and with dropout ResNet-50 model.	Link
Day to Night Domain Adaptation	Testing domain adaptation using ResNet-50 on the CIFAR dataset, where the source distribution will be the original images obtained during the day and the target distribution will be the transformed version of the dataset created by altering the brightness of the images to generate a night time dataset. Comparing the results with the L2 regularized ResNet-50 model.	Link