Towards convergence rate analysis of random forests for classification

W Gao, F Xu, ZH Zhou - Artificial Intelligence, 2022 - Elsevier
Random forests have been one of the successful ensemble algorithms in machine learning,
and the basic idea is to construct a large number of random trees individually and make …

[PDF][PDF] Adaptivity to noise parameters in nonparametric active learning

A Locatelli, A Carpentier… - Proceedings of the 2017 …, 2017 - proceedings.mlr.press
Adaptivity to Noise Parameters in Nonparametric Active Learning Page 1 Proceedings of
Machine Learning Research vol 65:1–34, 2017 Adaptivity to Noise Parameters in …

Recovery guarantees for polynomial coefficients from weakly dependent data with outliers

LST Ho, H Schaeffer, G Tran, R Ward - Journal of Approximation Theory, 2020 - Elsevier
Learning non-linear systems from noisy, limited, and/or dependent data is an important task
across various scientific fields including statistics, engineering, computer science …

Fast learning rates with heavy-tailed losses

VC Dinh, LS Ho, B Nguyen… - Advances in neural …, 2016 - proceedings.neurips.cc
We study fast learning rates when the losses are not necessarily bounded and may have a
distribution with heavy tails. To enable such analyses, we introduce two new conditions:(i) …

A generalization bound of deep neural networks for dependent data

QH Do, BT Nguyen, LST Ho - Statistics & Probability Letters, 2024 - Elsevier
Existing generalization bounds for deep neural networks require data to be independent
and identically distributed (iid). This assumption may not hold in real-life applications such …

Interclass interference suppression in multi-class problems

J Liu, M Bai, N Jiang, R Cheng, X Li, Y Wang, D Yu - Applied Sciences, 2021 - mdpi.com
Multi-classifiers are widely applied in many practical problems. But the features that can
significantly discriminate a certain class from others are often deleted in the feature selection …

A Generalization Bound of Deep Neural Networks for Dependent Data

QH Do, BT Nguyen, LST Ho - arXiv preprint arXiv:2310.05892, 2023 - arxiv.org
Existing generalization bounds for deep neural networks require data to be independent
and identically distributed (iid). This assumption may not hold in real-life applications such …

An adaptive multiclass nearest neighbor classifier

N Puchkin, V Spokoiny - ESAIM: Probability and Statistics, 2020 - esaim-ps.org
We consider a problem of multiclass classification, where the training sample S_n={(X i, Y i)}
ni= 1 is generated from the model ℙ (Y= m| X= x)= η m (x), 1≤ m≤ M, and η 1 (x),…, η M (x) …

Joint Dataset Reconstruction and Power Control for Distributed Training in D2D Edge Network

J Wu, J Wu, L Chen, Y Sun, Y Wu - IEEE Transactions on …, 2023 - ieeexplore.ieee.org
The intrinsic nature of non-independent and identically distributed datasets on
heterogeneous devices slows down the distributed model training process and reduces the …

Adaptive group Lasso neural network models for functions of few variables and time-dependent data

LST Ho, N Richardson, G Tran - Sampling Theory, Signal Processing, and …, 2023 - Springer
Learning nonlinear functions from time-varying measurements is always difficult due to the
high correlation among observations. This task is more challenging when the target function …