Transformer has achieved excellent performance in the knowledge tracing (KT) task, but they are criticized for the manually selected input features for fusion and the defect of single global context modelling to direc...
Transformer has achieved excellent performance in the knowledge tracing (KT) task, but they are criticized for the manually selected input features for fusion and the defect of single global context modelling to directly capture students' forgetting behavior in KT, when the related records are distant from the current record in terms of time. To address the issues, this paper first considers adding convolution operations to the Transformer to enhance its local context modelling ability used for students' forgetting behavior, then proposes an evolutionary neural architecture search approach to automate the input feature selection and automatically determine where to apply which operation for achieving the balancing of the local/global context modelling. In the search space design, the original global path containing the attention module in Transformer is replaced with the sum of a global path and a local path that could contain different convolutions, and the selection of input features is also considered. To search the best architecture, we employ an effective evolutionary algorithm to explore the search space and also suggest a search space reduction strategy to accelerate the convergence of the algorithm. Experimental results on the two largest and most challenging education datasets demonstrate the effectiveness of the architecture found by the proposed approach.
Dear editor,Polar codes, along with a low complexity successive cancellation(SC) decoding, were discovered in [1] by Ar?kan. It is shown in [2] that the bit errors with the SC decoding are correlated. To improve the p...
详细信息
Dear editor,Polar codes, along with a low complexity successive cancellation(SC) decoding, were discovered in [1] by Ar?kan. It is shown in [2] that the bit errors with the SC decoding are correlated. To improve the performance, other decoding techniques [3, 4] and concatenation schemes [5, 6] were studied. In this study, the theoretical aspects of the error correlation are investigated to improve the decoding performance of polar codes. This is the first attempt to utilize the error correlation to improve the performance of polar codes.
With the capability to sculpt complex radiative wavefronts and couple energy from the reference wave to the desired radiation pattern, metasurface antennas have gained tremendous progress in recent years. The flexibil...
详细信息
A wideband bidirectional circularly polarized (CP) orbital angular momentum (OAM) antenna array is proposed for millimeter-wave frequency. The designed antenna array is composed of four bidirectional radiating dual-CP...
详细信息
The authors study the binary codes spanned by the adjacency matrices of the strongly regular graphs(SRGs)on at most two hundred vertices whose existence is *** authors show that in length less than one hundred they ca...
详细信息
The authors study the binary codes spanned by the adjacency matrices of the strongly regular graphs(SRGs)on at most two hundred vertices whose existence is *** authors show that in length less than one hundred they cannot be cyclic,except for the exceptions of the SRGs of parameters(85,42,20,21)and(96,60,38,36).In particular,the adjacency code of a(85,42,20,21)is the zero-sum *** the range[100,200]the authors find 29 SRGs that could possibly have a cyclic adjacency code.
CNNs (Convolutional Neural Networks) have a good performance on most classification tasks, but they are vulnerable when meeting adversarial examples. Research and design of highly aggressive adversarial examples can h...
CNNs (Convolutional Neural Networks) have a good performance on most classification tasks, but they are vulnerable when meeting adversarial examples. Research and design of highly aggressive adversarial examples can help enhance the security and robustness of CNNs. The transferability of adversarial examples is still low in black-box settings. Therefore, an adversarial example method based on probability histogram equalization, namely HE-MI-FGSM (Histogram Equalization Momentum Iterative Fast Gradient Sign Method) is proposed. In each iteration of the adversarial example generation process, the original input image is randomly histogram equalized, and then the gradient is calculated to generate adversarial perturbations to mitigate overfitting in the adversarial example. The effectiveness of the method is verified on the ImageNet dataset. Compared with the advanced method I-FGSM (Iterative Fast Gradient Sign Method) and MI-FGSM (Momentum I-FGSM), the attack success rate in the adversarial training network increased by 27.9% and 7.7% on average, respectively.
This paper presents a hybrid method for determination of continuous dielectric properties of clothing materials. The dielectric constant and loss tangent of three types of materials are firstly investigated using open...
详细信息
Gathering reliable lab.led samples for polarimetric synthetic aperture (PolSAR) image classification is lab.rious. Moreover, applying a trained classifier to new domains often leads to noticeable performance degradati...
详细信息
In wireless communication systems, accurate channel estimation is essential to ensure the performance of wireless communication systems. Massive Multiple Input Multiple Output (M-MIMO) systems have a dramatic increase...
详细信息
暂无评论