ID 原文 译文
56748 通过全面的模拟数据实验和真实数据实验的分析结果,我们展示了MSS模型在模型预测和变量选择方面同时具备随机效应模型(random effects models)和稀疏回归模型(sparse regression models)的优势,相比已有方法大幅提高泛化性能. To overcome the computational complexity of Bayesian inference,an efficient algorithm is proposed based on variational inference, which is scalable to large-scale data analysisproblems.
56749 MSS模型通过对多任务学习模型中不同效应的区分,能够更加有效的识别模型中的共享随机效应和特异稀疏效应,进而增强模型在模型预测和变量选择方面的性能. The effectiveness of MSS in prediction and variable selection is demonstrated through comprehensivesimulation studies and real data analysis of movie rating. The results demonstrate that the characterization ofshared weak effects and task-specific sparse effects can improve the accuracy of prediction and variable selection
56750 数据类型和分布的复杂化导致样本间关系的不确定性增强,给有效挖掘数据的潜在类簇结构带来挑战. The complexity of data types and distributions leads to an increase in uncertainty of the relationshipsbetween data samples, which bring challenges in discovering the cluster structure inherent in a data set.
56751 为降低样本关系不确定性对数据聚类带来的影响,本文将聚类集成中样本稳定性概念扩展至聚类分析中. Toaddress this challenge, this paper presents the concept of the sample’s stability in a clustering ensemble, whichis extended to the area of clustering analysis.
56752 本文从理论上分析样本稳定的合理性,并提出基于信息熵的样本稳定性度量方法. We theoretically analyze the rationality of the sample’s stabilityand propose an entropy-based sample’s stability measure.
56753 此外,本文提出一个基于样本稳定性的聚类方法,该方法先将数据分为稳定样本集和不稳定样本集,然后挖掘稳定样本的团簇结构,并将不稳定样本划分至该团簇结构中. Besides, we propose a clustering method based on thesample’s stability. The proposed method divides a data set into stable and unstable samples, discovers the clusterstructure of the stable samples, and assigns the unstable samples into this structure.
56754 最后,本文通过二维人造数据和图像分割场景可视化显示样本稳定性的合理性,并在基准数据集上验证本文所提聚类算法的有效性. The results of experimentson two-dimensional data sets and an image segmentation data set visually demonstrate the rationality of thesample’s stability concept and effectiveness of the proposed clustering method based on the sample’s stabilitymeasure.
56755 公平性学习是机器学习领域的研究热点,预防歧视的目的在于执行预测任务之前消除不公平训练集对于分类器的影响. Fairness learning is one of research hotspots in machine learning. The purpose of preventing dis?crimination is to eliminate the impact of unfair training sets on classifiers before performing prediction tasks.
56756 为了保证分类公平性和准确性,本文通过发现和消除原始数据集中的歧视样本寻找生成公平数据集的方法,即提出了一种基于分类间隔的加权方法用于处理二分类任务中的歧视现象,并在demographic parity和equalized odds公平性判定准则上实现分类公平. To ensure the fairness and accuracy of classification, this paper presents a method for generating fair data setsby identifying and eliminating discriminatory samples in original data sets. This is a margin-based weightedmethod for dealing with discrimination in binary classification tasks and obtaining the demographic parity andequalized odds. To improve the classification accuracy, the target set is selected after projecting based on themargin principle.
56757 为了不影响分类准确性,本文基于最大间隔原理将样本投影之后选出目标集,对于目标集中的每个样本,通过加权距离度量方法判定该样本是否具有歧视性,并进行修正. For each sample in the target set, a weighted distance measurement method is used to identifythe discriminatory sample and then correct it.