Search

The fairness & bias of narrow coverage of AI becomes another challenge for AI researchers. If a commercial AI trains with a biased dataset, there will be severe gender or racial fairness and bias issues. Since the researchers use primary language datasets to train AI, the broad audience cannot be satisfied if a novel LLM (Large Language Model) AI shows a knowledge or creativity limitation on their specific spoken language. Narrow coverage of the LLMs can lead the audience to misinterpretation and confusion if the service involves STT (Speech-To-Text). In this paper, to overcome this issue of data diversity, we propose the idea that the embedded, extracted features have captured semantic proximity information that can be useful to mitigate diversity issues. This project focused on the Korean language food dataset for STT services, where a narrow-trained A.I. is prone to show its limitations, such as lifestyle-related elements. To present our proof of concept, we trained a baseline model, GPT2, with the Korean Wikipedia dataset in 2022. Then, we employed DistilBERT and KoBERT for comparison. The extracted hidden_state_output features from each model were utilized to build feature-extraction-based text search engines. We used the same idea of Local Sensitive Hashing (LSH) but effectively located a similar hash by applying transposed weights. We also present conventional classification benchmarks for performance comparison using top-k measurements, times for training and memory & disc consumptions. In the discussion, we proposed that our idea can mitigate the diversity problem without re-training the model and tokenizer.

This paper proposes a Design Space Exploration for Edge machine learning through the utilization of the novel MathWorks FPGA Deep Learning Processor IP, featured in the HDL Deep Learning toolbox. With the ever-increasing demand for real-time machine learning applications, there is a critical need for efficient and low-latency hardware solutions that can operate at the edge of the network, in close proximity to the data source. The HDL Deep Learning toolbox provides a flexible and customizable platform for deploying deep learning models on FPGAs, enabling effective inference acceleration for embedded IoT applications. In this study, our primary focus lies in investigating the impact of parallel processing elements on the performance and resource utilization of the FPGA-based processor. By analyzing the trade-offs between accuracy, speed, energy efficiency, and hardware resource utilization, we aim to gain valuable insights into making optimal design choices for FPGA-based implementations. Our evaluation is conducted on the AMD-Xilinx ZC706 development board, which serves as the target device for our experiments. We consider all the compatible Convolutional Neural Networks available within the HDL Deep Learning toolbox to comprehensively assess the performances.

The development of information technology has promoted the expansion of the application field of facial recognition technology. Its mainstream recognition methods rely on deep learning algorithms for calculation, but the problem of large data computation brought by its system makes it difficult to apply to embedded platform devices. As a result, this study focuses on improving recognition systems built on lightweight backend networks and builds an embedded platform system environment using multi-scale feature fusion, anchor box size optimization, the addition of channel attention mechanism weighted features, affine face alignment, and file compilation. The experimental results showed that when the number of iterations was 300, the loss value (0.46) of the improved embedded algorithm was much smaller than that of other comparison algorithms (1.42, 1.73, 2.01), and its ACC value (0.924) was significantly better than other comparison algorithms (0.915, 0.909, 0.894). The minimum system testing time consumedwas7 ms. This deep learning embedded facial recognition algorithm has high recognition accuracy, a fast running speed, and is less limited by environmental conditions and data types. It is ideally suited for use in embedded hardware devices, broadening the scope of equipment matching and facial recognition algorithms’ applications. As a result, it is better suited to satisfy the demands of embedded devices and massive data processing jobs.

The use of Convolutional Neural Networks (CNN) for the application of wood defects detection has gained significant attention in recent years. In industrial settings, these tasks are typically performed in a strict and consistent environment, making the use of large and complex CNN models unnecessary. Despite this, recent research has continued to focus on the use of such models to achieve increasingly accurate detections. These models require costly machines for inference, making adoption less likely especially for manufacturers in the developing nations. In view of this limitation, this paper proposes a set of strategies to achieve a highly efficient CNN model for fast and accurate wood defects detection based on the YOLOv4-Tiny architecture. The model has been improved with Efficient Channel Attention (ECA) to better select multi-scale features from the backbone network and has been drastically reduced in size through an iterative pruning and recovery process. This results in an 88% reduction in model parameters while retaining accuracy comparable to most state-of-the-art (SOTA) methods. Consequently, the model can perform near real-time inference directly on a general-purpose embedded processor without external hardware accelerators. This research hopes to motivate the development of efficient defect detectors that can run on low-cost embedded devices.

Virtualization has become one of the main tools for making efficient use of the resources offered by multicore embedded platforms. In recent years, even sectors such as space, aviation, and automotive, traditionally wary of adopting this type of technology due to the impact it could have on the safety of their systems, have been forced to introduce it into their day-to-day work, as their applications are becoming increasingly complex and demanding. This article provides a comprehensive review of the research work that uses or considers the use of a hypervisor as the basis for building a virtualized safety-critical embedded system. Once the hypervisors developed or adapted for this type of system have been identified, an exhaustive qualitative comparison is made between them. an exhaustive qualitative comparison is made between them. To the best of our knowledge, this is the first time that all this information is collected in a single article. Therefore, the main contribution of this article is that it collects and categorizes the information of each hypervisor and compares them with each other, so that this article can be used as a starting point for future researchers in this area, who will be able to quickly check which hypervisor is best suited to their research needs.

Rapier looms are currently important equipment in the weaving process. The control system of the loom determines the performance of the loom, to a large extent. To effectively reduce the production cost and energy consumption and to improve the start-up performance and production efficiency of rapier looms in industrial production, this paper develops an integrated rapier loom control system based on the direct drive of a switched reluctance motor (SRM) spindle and conducts field tests and applications. The contribution and innovation of this paper is to develop a complete set of low-cost control systems, propose an SRM single neuron fuzzy PID speed control strategy based on voltage chopping control and use it for the control of the main shaft drive technology of rapier looms. The integrated rapier loom control system based on the SRM spindle direct drive proposed in this paper reduces the production costs and energy consumption and improves the start-up performance and production efficiency of the rapier loom. This text carries on the systematic plan design to the control system from the hardware system and the software system. First, according to actual needs, starting from the aspect of reducing control costs and combined with the characteristics of embedded systems, such as tailorability, low cost, and strong scalability, this paper proposes a control system hardware structure based on CAN bus communication and a fully embedded STM32. The control system is divided into multiple control modules, such as the main control module, the spindle drive module, and the power transmission coil module. The system conducts a distributed control to the loom through the CAN bus and is equipped with various communication interfaces, such as Ethernet and RS485. Second, combining the characteristics of the SRM with a simple structure, a large starting torque and the operation mode of the loom, the basic control mode of SRM voltage chopping control is determined. To improve the efficiency and start-up performance of the speed control system, the SRM single neuron PID control algorithm is proposed, and a single neuron is used to improve the PID parameters. On this basis, fuzzy control is introduced to adjust the output gain of a single neuron PID control online to improve the system performance and reduce system energy consumption. Finally, the entire set of rapier loom control systems was verified, tested and debugged on site. The results show that each functional circuit works normally, and that the designed control system can meet the speed response demand of the loom at 850 rpm and reduce the production cost and energy consumption. The comparison experiment between the single neuron fuzzy PID algorithm of the motor and the traditional PID control algorithm in the actual loom production process proves that the proposed control algorithm has a better dynamic response performance. The proposed control algorithm effectively improves the starting performance and production efficiency of rapier looms and meets the actual needs of industrial applications.

Embedded software is the basic guarantee for the safe and stable operation of e-commerce platform. In order to improve the embedded software quality of e-commerce platform, a multi-objective enhancement method based on firefly algorithm optimization is proposed. Based on the two-dimensional strengthening model of embedded software duration cost of e-commerce platform, the application effect of embedded software is quantified by setting multi-level index system, so that the application effect of embedded software is included in the strengthening target. On this basis, the multi-objective strengthening model of embedded software duration, cost and application effect is constructed. The firefly algorithm is used to solve the multi-objective enhancement model, and the firefly algorithm is optimized by constraint priority non inferior ranking. The optimal compromise solution is selected according to the fuzzy membership degree in fuzzy mathematics. Experimental results show that the proposed method can significantly improve the quality of embedded software of e-commerce platform.

In view of the current situation, that physiological parameter monitoring systems can only achieve local monitoring, and the multi-physiological parameter monitors are large, expensive, and disadvantageous to remote monitoring. This paper combines embedded and mobile communication technologies to develop a new type of multi-physiological parameter medical monitoring system with remote data transmission function. First, through the analysis of embedded system principles, an embedded computer system based on ARM is designed. Secondly, the human-computer interaction interface, data acquisition, and analysis module are designed. Finally, by connecting to the Internet network to communicate with the medical center server, the remote transmission of local detection data and the issuance of alarm signals when dangerous situations occur are realized. The system can collect and display multiple physiological parameters such as heart rate, blood pressure, blood oxygen saturation, and body temperature in real time. The simulation experiment results show that the system's monitoring function and remote data transmission function meet the design requirements, can quickly and accurately find out-of-standard data, and perform remote alarm. The system is small, easy to expand, stable in data transmission, high in reliability, convenient for remote monitoring and data sharing, and is an ideal monitoring device for hospitals and community medical centers.

Traffic sign recognition is a very important computer vision task for a number of real-world applications such as intelligent transportation surveillance and analysis. While deep neural networks have been demonstrated in recent years to provide the state-of-the-art performance traffic sign recognition, a key challenge for enabling the widespread deployment of deep neural networks for embedded traffic sign recognition is the high computational and memory requirements of such networks. As a consequence, there are significant benefits in investigating compact deep neural network architectures for traffic sign recognition that are better suited for embedded devices. In this paper, we introduce MicronNet, a highly compact deep convolutional neural network for real-time embedded traffic sign recognition designed based on macroarchitecture design principles (e.g., spectral macroarchitecture augmentation, parameter precision optimization, etc.) as well as numerical microarchitecture optimization strategies. The resulting overall architecture of MicronNet is thus designed with as few parameters and computations as possible while maintaining recognition performance, leading to optimized information density of the proposed network. The resulting MicronNet possesses a model size of just 1 MB and 5 10000 parameters ( 27× fewer parameters than state-of-the-art) while still achieving a human performance level top-1 accuracy of 98.9% on the German traffic sign recognition benchmark. Furthermore, the MicronNet requires just 10 million multiply-accumulate operations to perform inference, and has a time-to-compute of just 32.19 ms on a Cortex-A53 high efficiency processor. These experimental results show that the highly compact, optimized deep neural network architectures can be designed for real-time traffic sign recognition that are well-suited for embedded scenarios.

Model-driven engineering has got a foothold in industry as an effective way to tame the complexity of modern software, which is meant to run on embedded systems with real-time constraints by promoting abstraction, in terms of prescriptive models, and automation, in terms of model manipulations. In the plethora of modeling languages, the unified modeling language (UML) has emerged and established itself as a de facto standard in industry, the most widely used architectural description language and an ISO/IEC standard. In the SMARTCore project, we have provided solutions for the UML-based development of software to run on multicore embedded real-time systems with the specific focus of automating the generation of executable code and the optimization of task allocation based on a unique combination of model-based and execution-based mechanisms. In this paper, we describe the lessons learned in the research work carried out within SMARTCore and provide a set of perspectives that we consider to be highly relevant for the forthcoming future of this research area to enable a wider adoption of UML-based development in industry in general, and in the multicore embedded real-time domain in particular.