The architecture information of a Deep Neural Network (DNN) model is considered a valuable, sensitive piece of property for a company. Knowledge of a DNN’s exact architecture allows any adversary to build a substitute model and use this substitute model to launch devastating adversarial attacks. Side-channel based DNN architecture stealing can...

­Background Recurrent neural networks (RNNs) that enable accurate automatic speech recognition (ASR) are large in size and have long short-term memory (LSTM) capabilities. Due to the large size of these networks, most speech recognition tasks are performed in the cloud servers, which requires constant internet connection, introduces privacy concerns,...

­Recently, Deep Neural Networks (DNNs) have been deployed in many safety-critical applications. The security of DNN models can be compromised by adversarial input examples, where the adversary maliciously crafts and adds input noise to fool a DNN model. The perturbation of model parameters (e.g., weight) is another security concern, one that relates...

Recent breakthroughs in deep neural networks (DNNs) have led to improvements in state-of-the-art speech applications. Conventional DNNs have hundreds or thousands of neurons in each layer, which require a large amount of memory to store the connections between neurons. Implementing these networks in hardware requires a large memory and high computation...