Context This data is encoded in some mysterious way. What you do is train a neural network to decode the data! Content The data set contain to files, voc_train.txt and voc_test.txt, which contain training set and test set. Each row of data contain tow part, the encoded data and origin data, separate by space. e.g."J+oq398rhPl5Jc0kUBoD6A== himself54". "himself54" is origin data, "J+oq398rhPl5Jc0kUBoD6A==" is encode data. You can try whatever you can to decode the dataset.