ABSTRACT
Mobile devices have become an essential part of our daily lives. By virtue of both their increasing computing power and the recent progress made in AI, mobile devices evolved to act as intelligent assistants in many tasks rather than a mere way of making phone calls. However, popular and commonly used tools and frameworks for machine intelligence are still lacking the ability to make proper use of the available heterogeneous computing resources on mobile devices. In this paper, we study the benefits of utilizing the heterogeneous (CPU and GPU) computing resources available on commodity android devices while running deep learning models. We leveraged the heterogeneous computing framework RenderScript to accelerate the execution of deep learning models on commodity Android devices. Our system is implemented as an extension to the popular open-source framework TensorFlow. By integrating our acceleration framework tightly into TensorFlow, machine learning engineers can now easily make benefit of the heterogeneous computing resources on mobile devices without the need of any extra tools. We evaluate our system on different android phones models to study the trade-offs of running different neural network operations on the GPU. We also compare the performance of running different models architectures such as convolutional and recurrent neural networks on CPU only vs using heterogeneous computing resources. Our result shows that although GPUs on the phones are capable of offering substantial performance gain in matrix multiplication on mobile devices. Therefore, models that involve multiplication of large matrices can run much faster (approx. 3 times faster in our experiments) due to GPU support.
ABSTRACT
We believe that each individual is unique, and that it is necessary for diagnosis purpose to have a distinctive combination of signals and data features that fits the personal health status. It is essential to develop mechanisms for reducing the amount of data that needs to be transferred (to mitigate the troublesome periodically recharging of a device) while maintaining diagnostic accuracy. Thus, the system should not uniformly compress the collected physiological data, but compress data in a personalized fashion that preserves the 'important' signal features for each individual such that it is enough to make the diagnosis with a required high confidence level. We present a diagnostic quality driven mechanism for remote ECG monitoring, which enables a notation of priorities encoded into the wave segments. The priority is specified by the diagnosis engine or medical experts and is dynamic and individual dependent. The system pre-processes the collected physiological information according to the assigned priority before delivering to the backend server. We demonstrate that the proposed approach provides accurate inference results while effectively compressing the data.
