Fingerprinting malware by its behavioral signature has been an attractive approach for malware detection due to the homogeneity of dynamic execution patterns across different variants of similar families. Although previous research works show reasonably good performance in dynamic detection using machine learning techniques on a large corpus of training sets, in many practical defence scenarios, decisions must be undertaken based on a scarce number of observable samples. This paper demonstrates the effectiveness of generative adversarial autoencoder for dynamic malware detection under outbreak situations where in most cases a single sample is available for training the machine learning algorithm to detect similar samples in the wild.