Designing a robust active queue management(RAQM)is mandatory to avoid congestion in networks with wireless access links,because transmission control protocol(TCP)can detect con-gestion after its occurrence in a commun...
详细信息
Designing a robust active queue management(RAQM)is mandatory to avoid congestion in networks with wireless access links,because transmission control protocol(TCP)can detect con-gestion after its occurrence in a communication network and wireless links suffer from bottleneck capacity variations caused by fading and packet error rate(PER)in the acknowledgement ***,the average window size cannot be measured explicitly from the output signal and input and state delay are imposed to a network,which complicate theRAQMdesign problem in nonlinear ***,the main contribution of this study is to design a robust observer based control procedure based on integral sliding mode protocol to estimate the average window size,to control congestion in a TCP/RAQM network and to compensate input and state ***-ulation results via professional simulator NS-2 and SIMULINK confirm that the procedure can effectively estimate the window size and can robustly avoid congestion.
In this work, we address two main shortcomings of transformer architectures: input corruption and rank collapse in their output representation. We unveil self-attention as an autonomous state-space model that inherent...
详细信息
In this work, we address two main shortcomings of transformer architectures: input corruption and rank collapse in their output representation. We unveil self-attention as an autonomous state-space model that inherently promotes smoothness in its solutions, leading to lower-rank outputs and diminished representation capacity. Moreover, the steady-state solution of the model is sensitive to input perturbations. We incorporate a Proportional-Integral-Derivative (PID) closed-loop feedback control system with a reference point into the model to improve robustness and representation capacity. This integration aims to preserve high-frequency details while bolstering model stability, rendering it more noise-resilient. The resulting controlled state-space model is theoretically proven robust and adept at addressing the rank collapse. Motivated by this control framework, we derive a novel class of transformers, PID-controlled Transformer (PIDformer), aimed at improving robustness and mitigating the rank-collapse issue inherent in softmax transformers. We empirically evaluate the model for advantages and robustness against baseline transformers across various practical tasks, including object classification, image segmentation, and language modeling. Copyright 2024 by the author(s)
Mobile edge computing is becoming one of the ubiquitous computing paradigms to support applications requiring low latency and high computing capability. FPGA-based reconfigurable accelerators have high energy efficien...
详细信息
This research examines the transmission dynamics of the Omicron variant of COVID-19 using SEIQIcRVW and SQIRV models,considering the delay in converting susceptible individuals into infected *** significant delays eve...
详细信息
This research examines the transmission dynamics of the Omicron variant of COVID-19 using SEIQIcRVW and SQIRV models,considering the delay in converting susceptible individuals into infected *** significant delays eventually resulted in the pandemic’s *** ensure the safety of the host population,this concept integrates quarantine and the COVID-19 *** investigate the stability of the proposed *** fundamental reproduction number influences stability *** to our findings,asymptomatic cases considerably impact the prevalence of Omicron infection in the *** real data of the Omicron variant from Chennai,Tamil Nadu,India,is used to validate the outputs.
Bremsstrahlung—the spontaneous emission of broadband radiation from free electrons that are deflected by atomic nuclei—contributes to the majority of X-rays emitted from X-ray tubes and used in applications ranging ...
详细信息
Bremsstrahlung—the spontaneous emission of broadband radiation from free electrons that are deflected by atomic nuclei—contributes to the majority of X-rays emitted from X-ray tubes and used in applications ranging from medical imaging to semiconductor chip ***,we show that the bremsstrahlung intensity can be enhanced significantly—by more than three orders of magnitude—through shaping the electron wavefunction to periodically overlap with atoms in crystalline ***,we show how to shape the bremsstrahlung X-ray emission pattern into arbitrary angular emission profiles for purposes such as unidirectionality and ***,we find that these enhancements and shaped emission profiles cannot be attributed solely to the spatial overlap between the electron probability distribution and the atomic centers,as predicted by the paraxial and nonrecoil theory for free electron light *** work highlights an unprecedented regime of free electron light emission where electron waveshaping provides multi-dimensional control over practical radiation processes like *** results pave the way towards greater versatility in table-top X-ray sources and improved fundamental understanding of quantum electron-light interactions.
Dear editor,With the developments of industrial automation in recent years,vehicle suspension systems have received a great deal of attention in industry and academia due to their critical role in the chassis performa...
详细信息
Dear editor,With the developments of industrial automation in recent years,vehicle suspension systems have received a great deal of attention in industry and academia due to their critical role in the chassis performance of vehicles[1].The suspension system is expected to guarantee the vehicle’s maneuverability and provide satisfactory ride comfort by absorbing the vibrations arising from the road surface excitations and ensuring road-holding capability and suspension *** by the desirable performance of the model reference adaptive control(MRAC)approach,various literature studies have investigated its performance in diverse linear and nonlinear practical systems[2].
Federated learning is an effective method to train a machine learning model without requiring to aggregate the potentially sensitive data of agents in a central server. However, the limited communication bandwidth, th...
详细信息
The security of IoT (Internet of Things) devices and the protection of sensitive information processed by these devices such as personal data, sensor values, process-related information is an important and difficult c...
详细信息
For communications subject to correlated channel effects, noise recycling has recently been shown to enhance channel capacity with receiver-side-only changes. Using a taped-out chip, in a hard-detection scenario with ...
详细信息
In computer vision applications like surveillance and remote sensing,to mention a few,deep learning has had considerable *** imaging still faces a number of difficulties,including intra-class similarity,a scarcity of ...
详细信息
In computer vision applications like surveillance and remote sensing,to mention a few,deep learning has had considerable *** imaging still faces a number of difficulties,including intra-class similarity,a scarcity of training data,and poor contrast skin lesions,notably in the case of skin *** optimisation-aided deep learningbased system is proposed for accurate multi-class skin lesion *** sequential procedures of the proposed system start with preprocessing and end with *** preprocessing step is where a hybrid contrast enhancement technique is initially proposed for lesion identification with healthy *** of flipping and rotating data,the outputs from the middle phases of the hybrid enhanced technique are employed for data augmentation in the next ***,two pre-trained deep learning models,MobileNetV2 and NasNet Mobile,are trained using deep transfer learning on the upgraded enriched ***,a dual-threshold serial approach is employed to obtain and combine the features of both *** next step was the variance-controlled Marine Predator methodology,which the authors proposed as a superior optimisation *** top features from the fused feature vector are classified using machine learning *** experimental strategy provided enhanced accuracy of 94.4%using the publicly available dataset ***,the proposed framework is evaluated compared to current approaches,with remarkable results.
暂无评论