Search

Intelligent mobile image sensing powered by deep learning analyzes images captured by cameras from mobile devices, such as smartphones or smartwatches. It supports numerous mobile applications, such as image classification, face recognition, and camera scene detection. Unfortunately, mobile devices often lack the resources necessary for deep learning, leading to increased inference latency and rapid battery consumption. Moreover, the inference accuracy may decline over time due to potential data drift. To address these issues, we introduce a new cost-efficient framework, called Corun, designed to simultaneously handle multiple inference queries and continual model retraining/fine-tuning of a pre-trained model on a single commodity GPU in an edge server to significantly improve the inference throughput, upholding the inference accuracy. The scheduling method of Corun undertakes offline profiling to find the maximum number of concurrent inferences that can be executed along with a retraining job on a single GPU without incurring an out-of-memory error or significantly increasing the latency. Our evaluation verifies the cost-effectiveness of Corun. The inference throughput provided by Corun scales with the number of concurrent inference queries. However, the latency of inference queries and the length of a retraining epoch increase at substantially lower rates. By concurrently processing multiple inference and retraining tasks on one GPU instead of using a separate GPU for each task, Corun could reduce the number of GPUs and cost required to deploy mobile image sensing applications based on deep learning at the edge. [ABSTRACT FROM AUTHOR]

Deep learning models have significantly improved object detection, which is essential for visual sensing. However, their increasing complexity results in higher latency and resource consumption, making real-time object detection challenging. In order to address the challenge, we propose a new lightweight filtering method called L-filter to predict empty video frames that include no object of interest (e.g., vehicles) with high accuracy via hybrid time series analysis. L-filter drops those frames deemed empty and conducts object detection for nonempty frames only, significantly enhancing the frame processing rate and scalability of real-time object detection. Our evaluation demonstrates that L-filter improves the frame processing rate by 31–47% for a single traffic video stream compared to three standalone state-of-the-art object detection models without L-filter. Additionally, L-filter significantly enhances scalability; it can process up to six concurrent video streams in one commodity GPU, supporting over 57 fps per stream, by working alongside the fastest object detection model among the three models. [ABSTRACT FROM AUTHOR]

Depression is a serious mood disorder that is under-recognized and under-treated. Recent advances in mobile/wearable technology and ML (machine learning) have provided opportunities to detect the depressed moods of participants in their daily lives with their consent. To support high-accuracy, ubiquitous detection of depressed mood, we propose HADD, which provides new capabilities. First, HADD supports multimodal data analysis in order to enhance the accuracy of ubiquitous depressed mood detection by analyzing not only objective sensor data, but also subjective EMA (ecological momentary assessment) data collected by using mobile devices. In addition, HADD improves upon the accuracy of state-of-the-art ML algorithms for depressed mood detection via effective feature selection, data augmentation, and two-stage outlier detection. In our evaluation, HADD significantly enhanced the accuracy of a comprehensive set of ML models for depressed mood detection. [ABSTRACT FROM AUTHOR]

Emerging applications of IoT (the Internet of Things), such as smart transportation, health, and energy, are envisioned to greatly enhance the societal infrastructure and quality of life of individuals. In such innovative IoT applications, cost-efficient real-time decision-making is critical to facilitate, for example, effective transportation management and healthcare. In this paper, we formally define real-time decision tasks in IoT, review cutting-edge approaches that aim to efficiently schedule real-time decision tasks to meet their timing and data freshness constraints, review state-of-the-art approaches for efficient sensor data analytics in IoT, and discuss future research directions. [ABSTRACT FROM AUTHOR]

CNNs (Convolutional Neural Networks) are becoming increasingly important for real-time applications, such as image classification in traffic control, visual surveillance, and smart manufacturing. It is challenging, however, to meet timing constraints of image processing tasks using CNNs due to their complexity. Performing dynamic trade-offs between the inference accuracy and time for image data analysis in CNNs is challenging too, since we observe that more complex CNNs that take longer to run even lead to lower accuracy in many cases by evaluating hundreds of CNN models in terms of time and accuracy using two popular data sets, MNIST and CIFAR-10. To address these challenges, we propose a new approach that (1) generates CNN models and analyzes their average inference time and accuracy for image classification, (2) stores a small subset of the CNNs with monotonic time and accuracy relationships offline, and (3) efficiently selects an effective CNN expected to support the highest possible accuracy among the stored CNNs subject to the remaining time to the deadline at run time. In our extensive evaluation, we verify that the CNNs derived by our approach are more flexible and cost-efficient than two baseline approaches. We verify that our approach can effectively build a compact set of CNNs and efficiently support systematic time vs. accuracy trade-offs, if necessary, to meet the user-specified timing and accuracy requirements. Moreover, the overhead of our approach is little/acceptable in terms of latency and memory consumption. [ABSTRACT FROM AUTHOR]

In emerging Cyber-Physical Systems (CPS), the demand for higher communication performance and enhanced wireless connectivity is increasing fast. To address the issue, in our recent work, we proposed a dynamic programming algorithm with polynomial time complexity for effective cross-layer downlink Scheduling and Resource Allocation (SRA) considering the channel and queue state, while supporting fairness. In this paper, we extend the SRA algorithm to consider 5G use-cases, namely enhanced Machine Type Communication (eMTC), Ultra-Reliable Low Latency Communication (URLLC) and enhanced Mobile BroadBand (eMBB). In a simulation study, we evaluate the performance of our SRA algorithm in comparison to an advanced greedy cross-layer algorithm for eMTC, URLLC and LTE (long-term evolution). For eMTC and URLLC, our SRA method outperforms the greedy approach by up to 17.24%, 18.1%, 2.5% and 1.5% in terms of average goodput, correlation impact, goodput fairness and delay fairness, respectively. In the case of LTE, our approach outperforms the greedy method by 60%, 2.6% and 1.6% in terms of goodput, goodput fairness and delay fairness compared with tested baseline. [ABSTRACT FROM AUTHOR]