3D object detection is an important task in autonomous driving to perceive the surroundings. Despite excellent performance, existing detectors lack robustness real-world corruptions caused by adverse weathers, sensor noises, etc., provoking concerns about safety and reliability of systems. To comprehensively rigorously benchmark corruption detectors, this paper we design 27 types common for both LiDAR camera inputs considering realworld scenarios. By synthesizing these on public datasets,...

3D object detection is an essential perception task in autonomous driving to understand the environments. The Bird's-Eye-View (BEV) representations have significantly improved performance of detectors with camera inputs on popular benchmarks. However, there still lacks a systematic understanding robustness these vision-dependent BEV models, which closely related safety systems. In this paper, we evaluate natural and adversarial various representative models under extensive settings, fully...

Modern Deep Neural Network (DNN) accelerators are equipped with increasingly larger on-chip buffers to provide more opportunities alleviate the severe DRAM bandwidth pressure. However, most existing research on buffer utilization still primarily focuses single-layer dataflow scheduling optimization. As grow large enough accommodate weights in networks, impact of optimization communication diminishes significantly. Therefore, developing new paradigms that fuse multiple layers fully leverage...

The level of mobile government services has become a crucial metric for modernizing government’s governance capabilities, and it plays significant role in promoting sustainable development. This study aims to explore the key factors influencing across 31 provinces China, along with their combinatory pathways. Using technology–organization–environment (TOE) framework employing fuzzy-set qualitative comparative analysis (fsQCA), this conducts an in-depth service levels these provinces,...

The electrochemical behaviors and corrosion resistance of the wrought Mg–Y–Zn based alloys with high Y/Zn mole ratio have been investigated in details. results show that are dependent on modified arrangement LPSO phase by adjusting ratios. Increasing not only greatly decreases size plates, but also leads to precipitation Mg24Y5 phase. rate increases from 7.4 mg·cm−2·day−1 11.3 increasing up 3. It should be attributed decreasing plates as cathodes, further hydrogen evolution kinetics. related...

Traditional watershed segmentation is sensitive to noise and can leads serious over-segmentation. In order overcome the shortcomings of traditional segmentation, this paper presented an improved image method. Firstly, morphological opening/closing reconstruction filter applied remove noise. Secondly, multi-scale structure elements are used calculate gradient. Furthermore, gradient modified by viscous operators which most irregular local minimums. After standard transform, region merging...

3D object detection is an essential perception task in autonomous driving to understand the environments. The Bird's-Eye-View (BEV) representations have significantly improved performance of detectors with camera inputs on popular benchmarks. However, there still lacks a systematic understanding robustness these vision-dependent BEV models, which closely related safety systems. In this paper, we evaluate natural and adversarial various representative models under extensive settings, fully...

The aim of this work was to enhance the levels fruity esters in spine grape (Vitis davidii Foёx) wine by goal-directed amino acid supplementation during fermentation. HPLC and GC-MS monitored acids esters, respectively, alcoholic fermentation Cabernet Sauvignon grape. also used determine extracellular metabolites precursors involved synthesis esters. Alanine, phenylalanine, isoleucine were less than those Sauvignon. Pearson correlation between profile ester content two systems indicated that...

The star sensor, while delivering superior accuracy in spacecraft attitude measurement, is susceptible to disturbances such as high-dynamic motion, stray light, and environmental factors. These disruptions can lead a substantial decline or even abnormal output, risking disorientation mission failure. To address this, we propose an optimization-based tightly-coupled scheme, augmenting under dynamic conditions fortifying resilience situations like lost-in-space. Our approach involves...

The study explores the impact of AI tools on daily tasks designers in corporate environments, with a focus creation and evaluation processes design briefs. Given ChatGPT's advanced natural language processing capabilities its potential to meet complex communication analysis needs work, this tool was selected investigate application designers' workflows. Through expert interviews, experimental testing, third-party evaluations, we collected analyzed data understand work processes. findings...

Chrysene is a four-ring polycyclic aromatic hydrocarbon, which has been demonstrated to induce adverse effects through the inducement of oxidative stress. However, possible mechanisms and potential pathways chrysene-induced damages are scarcely investigated. In present study, stress xenobiotic metabolism disorder were investigated in human hepatocytes. could cause cell viability reduction cellular reactive oxygen species accumulation dose-dependent manner from 10 400 ng/mL. Moreover, mRNA...

The SnakeCLEF2023 competition aims to the development of advanced algorithms for snake species identification through analysis images and accompanying metadata. This paper presents a method leveraging utilization both Modern CNN models strong data augmentation are utilized learn better representation images. To relieve challenge long-tailed distribution, seesaw loss is in our method. We also design light model calculate prior probabilities using metadata features extracted from CLIP post...

Memory is a critical design consideration in current data-intensive DNN accelerators, as it profoundly determines energy consumption, bandwidth requirements, and area costs. As structures become more complex, larger on-chip memory capacity required to reduce data movement overhead, but at the expense of silicon Some previous works have proposed memory-oriented optimizations, such different reuse layer fusion schemes. However, these methods are not general potent enough cope with various...

Study tour is part of the comprehensive practical course middle schools, organic combination in-school education and out-of-school education, an innovative move from knowledge transfer to quality in transformation China’s basic education. With landing “Double Reduction” policy 2021, market for study tours has also seen significant changes. Based on analysis opportunities challenges brought by “double reduction” policy, article re-exams fundamental significance tours, educational value, puts...

Silicon modulators, which play a crucial role in silicon photonics systems, are currently trending towards lower biases and improved bandwidth. The plasma dispersion effect of modulators highlights the importance carrier concentration improving performance. Common doping profiles have been optimized for high efficiency but may suffer from increased loss. Our horizontal S-shaped modulator improves with excellent VπL 0.77 V·cm low loss 10.9 dB/cm, small resistance capacitance enhancing...

Provincial mobile government service platforms, represented by ‘Zheliban’ and ‘Yueshengshi’, have transformed the traditional way governments provide public services to citizens. Maintaining user engagement with these platforms has become a critical challenge in promoting digitalization of services. Despite widespread adoption services, characteristics influencing citizens’ intentions for continued usage not received sufficient attention academic literature. This study, based on Information...