Kernel PCA for Out-of-Distribution Detection
Kun Fang, Qinghua Tao, Kexin Lv, Mingzhen He, Xiaolin Huang, Jie Yang
in NeurIPS, 2024
This paper provides a framework of KPCA for OoD detection.
Publications
You can also find my articles on my Google Scholar profile.
Accepted Papers
Random Fourier Features for Asymmetric Kernels
Mingzhen He, Fan He, Fanghui Liu, Xiaolin Huang
in Machine Learning, 2024
This work presents a unified framework for kernel approximation via random Fourier features.
Decentralized Kernel Ridge Regression Based on Data-Dependent Random Feature
Ruikai Yang, Fan He, Mingzhen He, Jie Yang, Xiaolin Huang
in TNNLS, 2024
This work proposes a new decentralized KRR algorithm that pursues consensus on decision functions, which allows great flexibility and well adapts data on nodes.
Diffusion representation for asymmetric kernels via magnetic transform
Mingzhen He, Fan He, Ruikai Yang, Xiaolin Huang
in NeurIPS, 2024
This method maps an asymmetric kernel to the complex-valued field, embedding the symmetric and skew-symmetric parts to module and phase respectively, which enables a new dimension reduction method for data endowed with an asymmetric similarity such as directed graphs and trophic networks.
Learning non-parametric kernels via matrix decomposition for logistic regression
Kaijie Wang, Fan He, Mingzhen He, Xiaolin Huang
in Pattern Recognition Letters, 2023
This work presents several low rank decompostion schemes to adjust the kernel Gram matrices, which improves the flexibility of kernels without using conplicated rank penalty.
Learning with Asymmetric Kernels: Least Squares and Feature Interpretation
Mingzhen He, Fan He, Lei Shi, Xiaolin Huang, Johan A. K. Suykens
in TPAMI, 2023
This paper addresses the asymmetric kernel-based learning in the framework of the least squares support vector machine, resulting in the first classification method that can utilize asymmetric kernels directly.
Global Search and Analysis for the Nonconvex Two-Level ℓ₁ Penalty
Fan He, Mingzhen He, Lei Shi, Xiaolin Huang
in TNNLS, 2022
This work presents the nonconvex two-level l1 penalty based on tis piecewise linear property and apply it to machine learning tasks.