On-Device Dynamic DNN Inference through Spatial Sparsity Exploitation

Extended-Reality Technologies: An Overview of Emerging Applications in Medical Education and Clinical Care.

Machine learning has surged in popularity in recent years, with deep neural networks (DNNs) becoming a cornerstone of applications such as autonomous driving, augmented reality, and natural language processing, owing to their remarkable accuracy and scalability. As the demand for greater portability and flexibility grows, there is a strong push to optimize these applications for resource-constrained edge devices like smartphones and IoT devices.However, deploying DNN models on such devices presents significant challenges, primarily due to the large size of these models, which is a key contributor to their high performance. For instance, the DeepSpeech 2 model outperforms its predecessor, DeepSpeech 1, but its size is eight times larger, encompassing approximately 68 million parameters. Similarly, other networks designed for various tasks, including convolutional neural networks like VGG and transformer-based networks like BERT and GPT, also exhibit substantial sizes. For example, the GPT-3 model, a predecessor to ChatGPT, consists of an astonishing 175 billion parameters. Consequently, fitting DNN models onto resource-limited edge devices is a formidable challenge. My thesis addresses these challenges with a three-part approach: * High-Performance DNN Accelerators: I will develop DNN accelerators that excel in high performance, low latency, and energy efficiency through an algorithm and hardware co-design across diverse FPGA architectures (Chapter 2). * Real-Time DNN Execution: I aim to achieve exceptional accuracy and real-time execution for various large-scale DNNs on off-the-shelf mobile devices, IoT devices, and other platforms (Chapter 3). * CompVQC Framework: I will explore a systematic framework, CompVQC, to reduce the circuit length of quantum neural networks (QNNs) through model compression, which can achieve up to a 20% improvement in accuracy. This reveals that compilation-aware compression can enhance the robustness of QNNs on near-term noisy quantum devices (Chapter 4). The field of AI has recently undergone a significant transformation, ushering in the era of generative AI, where models like ChatGPT and Stable Diffusion create realistic and creative content for a wide range of applications. As generative AI becomes more prevalent, it is crucial to adapt algorithm-hardware co-design strategies to meet the specific needs of these models, particularly on edge devices. In the final chapter (Chapter 5), I will introduce my ongoing and future work focused on activating generative AI on the edge.--Author's abstract

With Generation Z becoming the primary group of nursing learners and the increasing shortage of nursing education resources, augmented reality and mixed reality based on head-mounted displays are being used more and more in nursing education. However, the current application landscape and the proper usage of these concepts remain unclear. Therefore, this study aims to conduct a scoping review to explore the current applications of head-mounted display-based augmented reality and mixed reality in nursing education and to clarify the definitions and usage of the concepts of augmented reality and mixed reality, ultimately providing directions for future applications and research. Based on the five-stage framework and PRISMA-ScR guidelines, a comprehensive collection and summarization of evidence regarding the application of head-mounted display-based augmented reality and mixed reality in nursing education were conducted. The databases retrieved include CNKI, Wanfang Database, VIP, CBM, PubMed, Cochrane Library, Embase, CINAHL, Web of Science, Scopus, IEEE Xplore, ACM Digital Library, and Ei. The languages of the included literature are Chinese and English. The retrieval was up to October 2024. A total of 44 studies were included in this review, covering three types of head-mounted displays: immersive head-mounted displays, smart glasses, and smartphone-based head-mounted displays. The main application areas were skills training and knowledge acquisition. Most of the studies were feasibility studies, though they also included some efficacy studies and research on personal experiences. In addition, these studies often employed vague or inconsistent definitions of augmented reality and mixed reality. Despite various explorations in the application of head-mounted display-based augmented reality and mixed reality in nursing education, there is still room for improvement in the relevant theory and utilization of this technology. In the future, interventions should use the three dimensions (observation of reality, real - virtual interaction, and fidelity of virtuality) described in detail, rather than simply employing the concepts of augmented reality or mixed reality. Efforts should be concentrated on developing and implementing head-mounted displays combined with other technologies that boast enhanced performance and cost - effectiveness, and further validating their effectiveness.

Applications like augmented reality, smart surveillance, and autonomous driving have made real-time video processing on edge devices more and more necessary. However, because of the restricted processing resources, memory limitations, and latency requirements, installing deep neural networks (DNNs) on edge devices presents major hurdles. the scalability of DNN architectures designed with the goal of maximizing model size, processing speed, and accuracy for real-time video processing on edge devices. We examine a number of methods, such as lightweight network designs, quantization, and model pruning, to determine how well they can lower computational load without sacrificing speed. According to experimental results, DNNs can significantly reduce latency and power consumption through smart tuning, allowing for effective real-time processing on devices with restricted resources. route toward high-performance, scalable DNN models, offering valuable perspectives on realistic deployment tactics for edge-based video processing applications.

Deep neural networks (DNN) have become indispensable tools for intelligent applications today. The demand for deploying DNN on the edge devices increases dramatically. Unfortunately, it is challenging because the DNN inference is computation-intensive, but edge devices are always resource-constraint. Prior solutions attempted to address these challenges with collaboration between cloud and edge devices, but they do not take the inference request rate into account. However, the inference delay will increase dramatically while the request rate becomes higher.In this paper, we propose a scheme to dynamic partition DNN into two or three parts and distribute them at the edge and cloud, achieving the lowest delay with the change of request rate. The scheme selects the optimal partition points of DNN with a layer evaluation model (LEM) and a total delay prediction model (DPM) under different request rates. The experiments of distributed deploying AlexNet, VGG, NiN and ResNet DNN models on image classification dataset ImageNet show that the proposed scheme significantly reduces the total end-to-end latency by fully using both the edge and cloud resources. It reduces the inference delay by 1.3 to 1.6 times and improves the throughput 1.2 to 1.7 times compared to the state of art partition approach.

Deep Neural Networks (DNNs) have been successfully applied to various real-world machine learning applications. However, performing large DNN inference tasks in real-time remains a challenge due to its substantial computational costs. Recently, Spiking Neural Networks (SNNs) have emerged as an alternative way of processing DNN'fs task. Due to its eventbased, data-driven computation, SNN reduces both inference latency and complexity. With efficient conversion methods from traditional DNN, SNN exhibits similar accuracy, while leveraging many state-of-the-art network models and training methods. In this work, an efficient neuromorphic hardware architecture for image recognition task is presented. To preserve accuracy, the analog-to-spiking conversion algorithm is adopted. The system aims to minimize hardware area cost and power consumption, enabling neuromorphic hardware processing in edge devices. Simulation results have shown that, with the MNIST digit recognition task, the system has achieved $\times 20$ reduction in terms of core area cost compared to the state-of-the-art works, with an accuracy of 94.4%, core area of 15 $\mu m^{2}$ at a maximum frequency of 250 MHz.

The perceptual behavior of consumers on a product displayed in the market has a vital role in analyzing the importance given to that product. Therefore, various strategies have been developed to understand this consumer behavior in the selection of products. Immersive technologies like virtual, augmented, and mixed reality are among them. With the foremost feature of immersion in the virtual world and interaction of users with virtual objects, virtual reality, and augmented reality have unlocked their potential in research and a user-friendly tool for analyzing consumer behavior. In addition to these technologies, mixed reality also has a significant role in investigating consumer behavior. Studies on immersive technologies in food applications are vast, hence this review focuses on the applications of virtual, augmented, and mixed reality in the food selection behavior of consumers. The behavioral studies are elicited to develop new products based on consumer needs, to understand the shopping behavior in supermarkets for real-time usage, and to know the influence of emotions in a selection of products. The findings suggest that virtual, augmented, and mixed reality induce immersion of the users in food selection behavioral studies. Information on the technological advancements in the tools used for bringing immersion and interaction are discussed for its futuristic applications in food. Though immersive technology gives users a realistic virtual environment experience, its application in food systems is in the budding stage. More research on human response studies would contribute to its innovative and inevitable application in the future.

Efficient deployment of Deep Neural Networks (DNNs) on edge devices (i.e., FPGAs and mobile platforms) is very challenging, especially under a recent witness of the increasing DNN model size and complexity. Model compression strategies, including weight quantization and pruning, are widely recognized as effective approaches to significantly reduce computation and memory intensities, and have been implemented in many DNNs on edge devices. However, most state-of-the-art works focus on ad hoc optimizations, and there lacks a thorough study to comprehensively reveal the potentials and constraints of different edge devices when considering different compression strategies. In this article, we qualitatively and quantitatively compare the energy efficiency of FPGA-based and mobile-based DNN executions using mobile GPU and provide a detailed analysis. Based on the observations obtained from the analysis, we propose a unified optimization framework using block-based pruning to reduce the weight storage and accelerate the inference speed on mobile devices and FPGAs, achieving high hardware performance and energy-efficiency gain while maintaining accuracy.

Deep neural networks (DNNs) have seen resurgent attraction to be implemented in edge applications. However, such implementations are not easy to achieve because execution of DNNs often require more resources than those provided by individual edge devices. On the other hand, relying on model-level distribution methods to implement a DNN on connected edge devices leads to costly communication overheads. To utilize available in-the-edge resources with less communication overhead, we propose using edge-tailored models comprised of nearly-independent narrow DNNs, the inference of which are accelerated using small cost-efficient RISC-based engines. We implement these engines on PYNQ boards as a platform that mimics the limited resources of edge devices. We create the narrow DNNs based on the available resources of PYNQ boards, and allocate each narrow DNN to one engine, implemented in an FPGA. We compare the communication overhead of our implantation against the state-of-the-art model-level distribution methods.

The cloud-only and edge-computing approaches have recently been proposed to satisfy the requirements of complex neural networks. However, the cloud-only approach generates a latency challenge because of the high data volumes that must be sent to a centralized location in the cloud. Less-powerful edge computing resources require a compression model for computation reduction, which degrades the model trading accuracy. To address this challenge, deep neural network (DNN) partitioning has become a recent trend, with DNN models being sliced into head and tail portions executed at the mobile edge devices and cloud server, respectively. We propose Edgepipe, a novel partitioning method based on pipeline inference with an image stream to automatically partition DNN computation between the edge device and cloud server, thereby reducing the global latency and enhancing the system-wide real-time performance. This method adapts to various DNN architectures, hardware platforms, and networks. Here, when evaluated on a suite of five DNN applications, Edgepipe achieves average latency speedups of 1.241× and 1.154× over the cloud-only approach and the state-of-the-art approach known as “Neurosurgeon”, respectively.

DeepEdgeSoC: End-to-end deep learning framework for edge IoT devices

As technology advances, mobile devices have gradually turned into wearable devices. Furthermore, virtual reality (VR), augmented reality (AR), and mixed reality (MR) are being increasingly applied in medical fields such as medical education and training, surgical simulation, neurological rehabilitation, psychotherapy, and telemedicine. Research results demonstrate the ability of VR, AR, and MR to ameliorate the inconveniences that are often associated with traditional medical care, reduce incidents of medical malpractice caused by unskilled operations, and reduce the cost of medical education and training. What is more, the application of these technologies has enhanced the effectiveness of medical education and training, raised the level of diagnosis and treatment, improved the doctor-patient relationship, and boosted the efficiency of medical execution. The present study introduces VR, AR, and MR applications in medical practice and education with the aim of helping health professionals better understand the applications and use these technologies to improve the quality of medical care.

Survey on Deep Neural Networks in Speech and Vision Systems

The success of deep neural networks (DNNs) is heavily dependent on computational resources. While DNNs are often employed on cloud servers, there is a growing need to operate DNNs on edge devices. Edge devices are typically limited in their computational resources, yet, often multiple edge devices are deployed in the same environment and can reliably communicate with each other. In this work we propose to facilitate the application of DNNs on the edge by allowing multiple users to collaborate during inference to improve their accuracy. Our mechanism, coined edge ensembles, is based on having diverse predictors at each device, which form an ensemble of models during inference. To mitigate the communication overhead, the users share quantized features, and we propose a method for aggregating multiple decisions into a single inference rule. We analyze the latency induced by edge ensembles, showing that its performance improvement comes at the cost of a minor additional delay under common assumptions on the communication network. Our experiments demonstrate that collaborative inference via edge ensembles equipped with compact DNNs substantially improves the accuracy over having each user infer locally, and can outperform using a single centralized DNN larger than all the networks in the ensemble together.

Three-dimensional (3D) imaging and image reconstruction play a prominent role in the diagnosis, treatment planning, and post-therapeutic monitoring of patients with congenital heart disease. More interactive and realistic medical experiences take advantage of advanced visualization techniques like augmented, mixed, and virtual reality. Further, 3D printing is now used in medicine. All these technologies improve the understanding of the complex morphologies of congenital heart disease. In this review article, we describe the technical advantages and disadvantages of various advanced visualization techniques and their medical applications in the field of congenital heart disease. In addition, unresolved issues and future perspectives of these evolving techniques are described.