Efficient federated item similarity model for privacy-preserving recommendation

Previous federated recommender systems are based on traditional matrix factorization, which can improve personalized service but are vulnerable to gradient inference attacks. Most of them adopt model averaging to fit the data heterogeneity of federated recommender systems, requiring more training costs. To address privacy and efficiency, we propose an efficient federated item similarity model for the heterogeneous recommendation, called FedIS, which can train a global item-based collaborative filtering model to eliminate user feature dependencies. Specifically, we extend the neural item similarity model to the federated model, where each client only locally optimizes the shared item feature matrix. We then propose a fast-convergent federated aggregation method inspired by meta-learning to address heterogeneous user updates and accelerate the convergence of global training. Furthermore, we propose a two-stage perturbation method to protect both local training and transmission while reducing communication costs. Finally, extensive experiments on four real-world datasets validate that FedIS can provide more competitive performance on federated recommendations. Our proposed method also shows significant training efficiency with less performance degradation.     

智能型试卷自动生成系统中试卷难度控制技术的研究

本文给出了试题难度和试卷难度的一种评估指标体系,讨论了试题、试卷难度的控制方法,并给出了试卷难度的一种控制模型和实现思路,以及该难度控制技术在智能型试卷自动生成系统中的实现,为智能型试卷自动生成系统中试卷难度的控制提供了理论依据。     

基于IRT的自动组卷算法模型研究

基于项目反应理论(IRT),采用离差加权值最小和测验项目信息函数值最高两个目标规划相结合的办法,设计了一个新的自动组卷数学模型。该模型克服了经典测量理论(CTT)的种种局限,有效解决了计量指标——信息量与非计量指标——题型分布、知识点分布等约束条件计算量纲不同的问题,很好地体现了测试项目信息函数值最高的组卷目标。     

一种新型的水下探测目标自动识别系统

利用布里渊散射提出了一种基于人工神经网络的水下探测目标识别系统,并在神经网络中引入动量项和采用可变步长。经过计算机仿真实验,结果表明：与传统的水下探测目标识别系统相比,该识别系统在识别时间和识别率两方面都得到了较大的改善．这些优点对实现水下探测目标的实时自动识别是极其有利的。     

高中学业水平考试研究(二):考试质量评价

学业水平考试试题和试卷的质量直接影响学业评价和诊断结果的有效性和可靠性。本文以上海市高中思想政治学科学业水平考试为例,从试题和题组功能偏差、试题得分与总分的相关系数、识别指数分析及分类一致性和准确性四个方面对考试的质量进行了定量评价,以介绍学业水平考试质量评价的方法。     

项目反应理论测验信度及其研究述评

项目反应理论下的测验信度能够评价潜在特质估计的可靠性与稳定性,由于具有宏观性的特点,项目反应理论信度的作用并不能被测验信息函数所取代,是IRT测验的一个重要指标。本文参考国内外文献,首先介绍国内外学者关于IRT信度作用的观点,并介绍和评价了多种IRT信度估计方法,然后简要介绍IRT信度的影响因素,最后展望了IRT信度领域后续研究尚可着力之处。     

Addressing the Lack of Measurement Invariance for the Measure of Acceptance of the Theory of Evolution

The Measure of Acceptance of the Theory of Evolution (MATE) was constructed to be a single-factor instrument that assesses an individual's overall acceptance of evolutionary theory. The MATE was validated and the scores resulting from the MATE were found to be reliable for the population of inservice high school biology teachers. However, many studies have utilized the MATE for different populations, such as university students enrolled in a biology or genetics course, high school students, and preservice teachers. This is problematic because the dimensionality and reliability of the MATE may not be consistent across populations. It is not uncommon in science education research to find examples where scales are applied to novel populations without proper assessment of the validity and reliability. In order to illustrate this issue, a case study is presented where the dimensionality of the MATE is evaluated for a population of non-science major preservice elementary teachers. With this objective in mind, factor analytic and item response models are fit to the observed data to provide evidence for or against a one-dimensional latent structure and to detect which items do not conform to the theoretical construct for this population. The results of this study call into question any findings and conclusions made using the MATE for a Hispanic population of preservice teachers and point out the error of assuming invariance across substantively different populations.     

项目教学法在传感器课程改革中的应用

本文针对传统《传感器技术及应用》课程教学中存在的不足,结合目前企业的实际需要,对课程进行了一些有益的探索和尝试;采用"项目教学法"的教学模式,将课程全部放在实验室进行,并对学生量身定制了"实验周"开展教学;本文通过实例介绍了项目教学法在传感器课程教学中的具体应用。     

“马尔可夫链蒙特卡洛”(MCMC)方法在估计IRT模型参数中的应用

本文介绍和阐述怎样运用“马尔可夫链蒙特卡洛”(MCMC)技术,并结合Bayes方法来估计IRT的模型参数。首先简要地概述了MCMC方法估计模型参数的基本原理;其次介绍MCMC方法估计模型参数的一般方法,涉及Gibbs抽样、取舍抽样、Metropolis-Hastings算法等概念和方法;最后以IRT的“二参数逻辑斯蒂”(2PL)模型为例,重点介绍了用“Gibbs范围内的M-H算法”估计项目参数(β1jβ2j)的算法过程。结束本文时还解说了MCMC方法的特点。阅读本文需具有随机过程、Markov链、Bayes方法等概率论的基本知识。