Elastic-DF: Scaling Performance of DNN Inference in FPGA Clouds through Automatic Partitioning

Customized compute acceleration in the datacenter is key to the wider roll-out of applications based on deep neural network (DNN) inference. In this article, we investigate how to maximize the performance and scalability of field-programmable gate array (FPGA)-based pipeline dataflow DNN inference accelerators (DFAs) automatically on computing infrastructures consisting of multi-die, network-connected FPGAs. We present Elastic-DF, a novel resource partitioning tool and associated FPGA runtime infrastructure that integrates with the DNN compiler FINN. Elastic-DF allocates FPGA resources to DNN layers and layers to individual FPGA dies to maximize the total performance of the multi-FPGA system. In the resulting Elastic-DF mapping, the accelerator may be instantiated multiple times, and each instance may be segmented across multiple FPGAs transparently, whereby the segments communicate peer-to-peer through 100 Gbps Ethernet FPGA infrastructure, without host involvement. When applied to ResNet-50, Elastic-DF provides a 44% latency decrease on Alveo U280. For MobileNetV1 on Alveo U200 and U280, Elastic-DF enables a 78% throughput increase, eliminating the performance difference between these cards and the larger Alveo U250. Elastic-DF also increases operating frequency in all our experiments, on average by over 20%. Elastic-DF therefore increases performance portability between different sizes of FPGA and increases the critical throughput per cost metric of datacenter inference.

Constructing software countermeasures against instruction manipulation attacks: an approach based on vulnerability evaluation using fault simulator

Fault injection attacks (FIA), which cause information leakage by injecting intentional faults into the data or operations of devices, are one of the most powerful methods compromising the security of confidential data stored on these devices. Previous studies related to FIA report that attackers can skip instructions running on many devices through many means of fault injection. Most existing anti-FIA countermeasures on software are designed to secure against instruction skip (IS). On the other hand, recent studies report that attackers can use laser fault injection to manipulate instructions running on devices as they want. Although the previous studies have shown that instruction manipulation (IM) could attack the existing countermeasures against IS, no effective countermeasures against IM have been proposed. This paper is the first work tackling this problem, aiming to construct software-based countermeasures against IM faults. Evaluating program vulnerabilities to IM faults is required to consider countermeasures against IM faults. We propose three IM simulation environments for that aim and compare them to reveal their performance difference. GDB (GNU debugger)-based simulator that we newly propose in this paper outperforms the QEMU-based simulator that we presented in AICCSA:1–8, 2020 in advance, in terms of evaluation time at most $$\times$$ × 400 faster. Evaluating a target program using the proposed IM simulators reveals that the IM faults leading to attack successes are classified into four classes. We propose secure coding techniques as countermeasures against IMs of each four classes and show the effectiveness of the countermeasures using the IM simulators.

A Comparative Study of Semantic Segmentation Models for Identification of Grape with Different Varieties

Accurate fruit segmentation in images is the prerequisite and key step for precision agriculture. In this article, aiming at the segmentation of grape cluster with different varieties, 3 state-of-the-art semantic segmentation networks, i.e., Fully Convolutional Network (FCN), U-Net, and DeepLabv3+ applied on six different datasets were studied. We investigated: (1) the segmentation performance difference of the 3 studied networks; (2) The impact of different input representations on segmentation performance; (3) The effect of image enhancement method to improve the poor illumination of images and further improve the segmentation performance; (4) The impact of the distance between grape clusters and camera on segmentation performance. The experiment results show that compared with FCN and U-Net the DeepLabv3+ combined with transfer learning is more suitable for the task with an intersection over union (IoU) of 84.26%. Five different input representations, namely RGB, HSV, L*a*b, HHH, and YCrCb obtained different IoU, ranging from 81.5% to 88.44%. Among them, the L*a*b got the highest IoU. Besides, the adopted Histogram Equalization (HE) image enhancement method could improve the model’s robustness against poor illumination conditions. Through the HE preprocessing, the IoU of the enhanced dataset increased by 3.88%, from 84.26% to 88.14%. The distance between the target and camera also affects the segmentation performance, no matter in which dataset, the closer the distance, the better the segmentation performance was. In a word, the conclusion of this research provides some meaningful suggestions for the study of grape or other fruit segmentation.

Effect of Character Marketing and Marketing Mix on Usage Intention of Internet-Only Banks: Evidence From South Korea

Two South Korean internet-only banks were established around the same time, but their performances differ significantly when the product is similar to intangible values such as brand equity and marketing efforts. This study credits brand equity with causing performance difference and proposes a research model to examine which marketing activities construct brand equity. Traditional 4P marketing activities and character marketing are selected as marketing mix activities. Additionally, this research examines whether marketing activity effectiveness is affected by gender and evaluates customer perception about internet-only banking based on a survey. The results show that all the considered marketing activities are significant in establishing brand equity; however, their effect on brand equity is moderated by gender. This research contributes to marketing by considering brand equity in the internet-only banking context and emphasizing the effect of marketing activities using characters. Industry managers could use these results to achieve competitiveness by fostering brand equity.

Soybean varieties tolerance to intercropping with maize

There is no available information in determining soybean varieties that shows tolerant when intercropped with maize. The objective of this study was to determine soybean varieties suitable for intercropping, based on changes in plant growth and yield when intercropped. The field experiment was conducted from October 2019 to February 2020 based on RCBD with 2 factors. The first factor was soybean with 16 national varieties. The second was cropping systems i.e., monoculture and intercropping. Morphology and yield data were observed and subsequently analysed based on performance difference (D) of each parameter between intercropping and monoculture. Tolerance level to cropping systems with maize was grouped based on the yield-based and selection index calculations. The results showed that different cropping systems affected morphological and yield parameters on 16 tested soybean varieties. Seven varieties i.e., Derap 1, Devon 1, Devon 2, Demas 1, Dena 1, Dena 2, and Wilis were grouped as tolerant to intercropping with maize. Among these, Derap 1 was the most tolerant variety.

Development of Eco-Concrete Produced with Calcined Marl Clay as Cement Replacement Material

Because of the increase in greenhouse gas emissions and the enforcement of environmental regulations, efforts to reduce CO2 emissions continue. In this study, calcined marl clay was used as a cement replacement material to minimize CO2 emissions by reducing the cement amount used. The raw materials for marl clay were analyzed, and the durability and chemical resistance of concrete produced with marl clay were examined. From the test results, the optimum replacement ratio of calcined marl clay was determined, and the performance difference with ordinary Portland cement (OPC) concrete was identified. The performance of the eco-concrete was comparable to that of the OPC concrete containing marl clay when the cement was optimally replaced with marl clay. Furthermore, the CO2 emission for each specimen was calculated. This study demonstrated the feasibility of developing eco-friendly concrete using calcined marl clay.

Constructing Automatic Classification Models for Chinese-language Chief Complaint (Preprint)

BACKGROUND Chief complaint is the initial, general, and written description of a patient’s symptoms provided during the hospital intake process. By improving the automatic classification of chief complaint text, the quality and efficiency of patients’ hospital visits can be improved. OBJECTIVE Using chief complaint data in Chinese from the Information Centre of Jiangsu Commission Health, we built models for automatically detecting the correct treating department and then conducted various tests on those models using machine learning and deep learning. METHODS The study tested and compared the performances of the traditional machine learning model of SVM with deep learning models of Bi-LSTM, Bi-LSTM-CRF, At-Bi-LSTM-CRF and Bi-GRU-CRF on the chief complaint text data mainly. It is mainly based on Chinese character expansion model train and test in all traditional machine learning and deep learning models. RESULTS We found that the Bi-LSTM performed better at the chief complaint classification task than the SVM and that the performance difference between the deep learning models constructed is not obvious. The F scores of Bi-LSTM, Bi-LSTM-CRF, At-Bi-LSTM-CRF and Bi-GRU-CRF model built for the experiment effectively reach 88.10, 87.91, 88.14 and 87.98. CONCLUSIONS We found that the Bi-LSTM performed better at the chief complaint classification task than the SVM and that the performance difference between the deep learning models constructed is not obvious. The F scores of Bi-LSTM, Bi-LSTM-CRF, At-Bi-LSTM-CRF and Bi-GRU-CRF model built for the experiment effectively reach 88.10, 87.91, 88.14 and 87.98.

Effect Of Rear Leg Morphology On Energetics Of A Quadrupedal Robot

Many biomimetic legged robots exist, but their leg designs appear to be arbitrarily chosen. Here, we examine the performance difference between a canine-inspired rear leg in its normal configuration versus the same leg in a transverse-mirrored configuration. A quadrupedal robot was built to test this hypothesis; the robot was successfully able to walk in with both rear-leg configurations. Successful telemetry of energy and localization data was also demonstrated. Both experimental and simulation results confirm that the transverse-mirrored configuration is faster and more efficient. In experiment the robot achieved speeds of up to 0.4 m/s versus 0.33 m/s, and specific resistances of 3.9 versus 5.1 in transverse and normal experiments, respectively. It is suggested here that the transverse-mirrored configuration, which engages the knee spring more than the normal configuration, be used in designs which require higher speeds and greater efficiencies.

Effect Of Rear Leg Morphology On Energetics Of A Quadrupedal Robot

Many biomimetic legged robots exist, but their leg designs appear to be arbitrarily chosen. Here, we examine the performance difference between a canine-inspired rear leg in its normal configuration versus the same leg in a transverse-mirrored configuration. A quadrupedal robot was built to test this hypothesis; the robot was successfully able to walk in with both rear-leg configurations. Successful telemetry of energy and localization data was also demonstrated. Both experimental and simulation results confirm that the transverse-mirrored configuration is faster and more efficient. In experiment the robot achieved speeds of up to 0.4 m/s versus 0.33 m/s, and specific resistances of 3.9 versus 5.1 in transverse and normal experiments, respectively. It is suggested here that the transverse-mirrored configuration, which engages the knee spring more than the normal configuration, be used in designs which require higher speeds and greater efficiencies.

Integrated Bimodal Fitting for Unilateral CI Users with Residual Contralateral Hearing

Background: The aim of this study was to compare, in users of bimodal cochlear implants, the performance obtained using their own hearing aids (adjusted with the standard NAL-NL1 fitting formula) with the performance using the Phonak Naìda Link Ultra Power hearing aid adjusted with both NAL-NL1 and a new bimodal system (Adaptive Phonak Digital Bimodal (APDB)) developed by Advanced Bionics and Phonak Corporations. Methods: Eleven bimodal users (Naìda CI Q70 + contralateral hearing aid) were enrolled in our study. The users’ own hearing aids were replaced with the Phonak Naìda Link Ultra Power and fitted following the new formula. Speech intelligibility was assessed in quiet and noisy conditions, and comparisons were made with the results obtained with the users’ previous hearing aids and with the Naída Link hearing aids fitted with the NAL-NL1 generic prescription formula. Results: Using Phonak Naìda Link Ultra Power hearing aids with the Adaptive Phonak Digital Bimodal fitting formula, performance was significantly better than that with the users’ own rehabilitation systems, especially in challenging hearing situations for all analyzed subjects. Conclusions: Speech intelligibility tests in quiet settings did not reveal a significant difference in performance between the new fitting formula and NAL-NL1 fittings (using the Naída Link hearing aids), whereas the performance difference between the two fittings was very significant in noisy test conditions.