Showing papers by "Veljko Pejovic published in 2023"
TL;DR: Mobiprox as mentioned in this paper implements tunable approximations of tensor operations and enables runtime adaptation of individual network layers, which can save up to 15% system-wide energy with a minimal impact on the inference accuracy.
Abstract: Runtime-tunable context-dependent network compression would make mobile deep learning adaptable to often varying resource availability, input"difficulty", or user needs. The existing compression techniques significantly reduce the memory, processing, and energy tax of deep learning, yet, the resulting models tend to be permanently impaired, sacrificing the inference power for reduced resource usage. The existing tunable compression approaches, on the other hand, require expensive re-training, seldom provide mobile-ready implementations, and do not support arbitrary strategies for adapting the compression. In this paper we present Mobiprox, a framework enabling flexible-accuracy on-device deep learning. Mobiprox implements tunable approximations of tensor operations and enables runtime adaptation of individual network layers. A profiler and a tuner included with Mobiprox identify the most promising neural network approximation configurations leading to the desired inference quality with the minimal use of resources. Furthermore, we develop control strategies that depending on contextual factors, such as the input data difficulty, dynamically adjust the approximation level of a model. We implement Mobiprox in Android OS and through experiments in diverse mobile domains, including human activity recognition and spoken keyword detection, demonstrate that it can save up to 15% system-wide energy with a minimal impact on the inference accuracy.
TL;DR: In this article , the authors discuss the challenges of ubiquitous computing in the "core" front -the computing itself, which appears to be rather orthodox and limited in comparison with the advancements on the "Core" front.
Abstract: Paving the way towards the realization of Mark Weiser's vision of ubiquitous computing [1], the research community has made incredible advancements on several fronts. When it comes to interacting with humans, for example, computers can already use pretty much anything as a touchpad [2]. Similarly, when it comes to sensing the environment, computers can unobtrusively detect anything from a driver fatigue [3] to the presence of the queen bee in a hive [4]. When compared with these, advancements on the "core" front - the computing itself - appear to be rather orthodox and limited.