For this project Vanya Cohen and I attempted to replicate Koch, Zemel, and Salakhutdinov’s paper Siamese Neural Networks for One-shot Image Recognition. The paper uses a convolutional siamese neural net to train a model in order to perform one shot learning on the omniglot dataset. The paper reports an accuracy of 92%. Our recreation managed to achieve an accuracy of 55%, mainly due to the constraints on our computational resources for training models and refining parameters.