 This paper examines the potential of machine learning methods, including artificial neural network, ANN, support vector machines, SVM, and random forest, RF, and different deep learning convolution neural networks, CNNs, for landslide detection. It compares these methods with the results of extensive fieldwork in the highly landslide-prone Razzawa district in Nepal. The authors found that the best result was achieved by a small window-sized CNN, which used spectral information only. Additionally, the addition of a 5-meter digital elevation model helped to distinguish between human settlements and landslides, but did not improve the overall classification accuracy. The authors concluded that CNNs have great potential for landslide mapping, but must be designed carefully and optimized for specific applications. This article was authored by Omid Gorbanzadeh, Thomas Blashk, Khalil Golamnya, and others.