In machine learning, the exponential growth of data and the associated ``curse of dimensionality'' pose significant challenges, particularly with expansive yet sparse datasets. Addressing these challenges, multi-view ensemble learning (MEL) has emerged as a transformative approach, with feature partitioning (FP) playing a pivotal role in constructing artificial views for MEL. Our study introduces the Semantic-Preserving Feature Partitioning (SPFP) algorithm, a novel method grounded in information theory. The SPFP algorithm effectively partitions datasets into multiple semantically consistent views, enhancing the MEL process. Through extensive experiments on eight real-world datasets, ranging from high-dimensional with limited instances to low-dimensional with high instances, our method demonstrates notable efficacy. It maintains model accuracy while significantly improving uncertainty measures in scenarios where high generalization performance is achievable. Conversely, it retains uncertainty metrics while enhancing accuracy where high generalization accuracy is less attainable. An effect size analysis further reveals that the SPFP algorithm outperforms benchmark models by large effect size and reduces computational demands through effective dimensionality reduction. The substantial effect sizes observed in most experiments underscore the algorithm's significant improvements in model performance.
A Novel Ontology-guided Attribute Partitioning Ensemble Learning Model for Early Prediction of Cognitive Deficits using Quantitative Structural MRI in Very Preterm Infants
Cincinnati Children's Hospital Medical Center(辛辛那提儿童医院医疗中心);University of Cincinnati College of Medicine(辛辛那提大学医学院);University of Cincinnati(辛辛那提大学)
IMOVNO+: A Regional Partitioning and Meta-Heuristic Ensemble Framework for Imbalanced Multi-Class Learning
Key Laboratory of Xinjiang Coal Resources Green Mining, Ministry of Education(新疆煤炭资源绿色开采重点实验室,教育部);Department of Mathematics, Ibn Khaldoun University(数学系,伊本·赫勒敦大学);LIM Laboratory of Informatics and Mathematics, Ibn Khaldoun University(信息与数学实验室,伊本·赫勒敦大学);School of Computing, Dublin City University(计算学院,都柏林城市大学)
Hydra: Preserving Ensemble Diversity for Model Distillation
Imperial College London(帝国理工学院);University of Amsterdam(阿姆斯特丹大学);ETH Zurich(苏黎世联邦理工学院);University of Warsaw(华沙大学);Google Research(谷歌研究院);University of California, Irvine(加州大学欧文分校)
Guided Random Forest and its application to data approximation
Department of Electrical Engineering, Indian Institute of Technology Delhi(印度理工学院德里分校电子工程系);Center for Computational Biology, Indraprastha Institute of Information Technology(印度信息科技学院计算生物学中心);Department of Computer Science and Engineering, Indraprastha Institute of Information Technology(印度信息科技学院计算机科学与工程系);Infosys Center for Artificial Intelligence, Indraprastha Institute of Information Technology(印度信息科技学院人工智能中心);Institute of Health and Biomedical Innovation, Queensland University of Technology(昆士兰科技大学健康与生物医学创新研究所);Department of Mathematics, Indian Institute of Technology Delhi(印度理工学院德里分校数学系)
FedCL-Ensemble Learning: A Framework of Federated Continual Learning with Ensemble Transfer Learning Enhanced for Alzheimer's MRI Classifications while Preserving Privacy
Privacy-Preserving Training of Tree Ensembles over Continuous Data
University of Washington Tacoma(华盛顿大学塔科马分校);Ghent University(根特大学)
大多数现有的安全多方计算(MPC)协议的隐私保护训练的决策树分布式数据假设的功能是分类的。在实际应用中,特征通常是数值。在具有连续值的数据上生长决策树的标准"in the clear“算法需要对每个特征的训练样本进行排序,以寻求每个节点中特征值范围内的最佳分割点。排序在MPC中是一个昂贵的操作,因此找到避免这种昂贵步骤的安全协议是隐私保护机器学习中的一个相关问题。在本文中,我们提出了三种更有效的方案,用于在具有连续特征的数据上安全地训练基于决策树的模型,即:(1)数据的安全离散化,然后在离散化数据上安全地训练决策树;(2)数据的安全离散化,然后在离散化数据上安全地训练随机森林;以及(3)在原始数据上的极端随机树("额外树“)的安全训练。方法(2)和(3)都涉及随机化特征选择。此外,在方法(3)中,也随机选择切割点,从而减轻了预先对数据进行排序或离散化的需要。我们实现了所有提出的解决方案,在半诚实设置与添加剂秘密共享的MPC。除了在数学上证明所有提出的方法都是正确和安全的之外,我们还在分类精度和运行时间方面进行了实验评估和比较。我们私下训练树集成在数据集上与1000个实例或功能在几分钟内,其准确性是在那些在明确获得的。这使得我们的解决方案比现有的基于不经意排序的方法更有效。
Addressing the Current Challenges of Quantum Machine Learning through Multi-Chip Ensembles
Interdisciplinary Program in Artificial Intelligence(人工智能交叉学科项目);Seoul National University(首尔国立大学);Department of Psychology(心理学系);Department of Brain and Cognitive Sciences(脑科学与认知科学系);Wells Fargo New York, USA(沃尔斯大通纽约美国分公司);Computational Science Initiative(计算科学倡议);Brookhaven National Laboratory(布鲁克海文国家实验室)