Recurrent Neural Networks (RNNs) with Long Short-Term Memory units (LSTM) are widely used because they are expressive and are easy to train.
This thesis report studies methods to solve Visual Question-Answering (VQA) tasks with a Deep Learning framework.
Dynamic analysis enables detecting Windows malware by executing programs in a controlled environment, and storing their actions in log reports.