MobiRNN: Efficient Recurrent Neural Network Execution on Mobile GPU
This work addresses privacy and efficiency issues for mobile applications using RNNs, but it is incremental as it builds on existing mobile deep learning optimizations by focusing on RNNs.
The paper tackled the problem of running Recurrent Neural Network (RNN) models locally on mobile devices to address privacy and efficiency concerns, and the result was MobiRNN, a mobile-specific optimization framework that significantly decreases latency, as shown in evaluations for activity recognition.
In this paper, we explore optimizations to run Recurrent Neural Network (RNN) models locally on mobile devices. RNN models are widely used for Natural Language Processing, Machine Translation, and other tasks. However, existing mobile applications that use RNN models do so on the cloud. To address privacy and efficiency concerns, we show how RNN models can be run locally on mobile devices. Existing work on porting deep learning models to mobile devices focus on Convolution Neural Networks (CNNs) and cannot be applied directly to RNN models. In response, we present MobiRNN, a mobile-specific optimization framework that implements GPU offloading specifically for mobile GPUs. Evaluations using an RNN model for activity recognition shows that MobiRNN does significantly decrease the latency of running RNN models on phones.