paper-with-me

홈 › Papers

Learning Sparse Visual Representations with Leaky Capped Norm Regularizers

2017-11-08 · Jianqiao Wangni, Dahua Lin

Sparsity inducing regularization is an important part for learning over-complete visual representations. Despite the popularity of $\ell_1$ regularization, in this paper, we investigate the usage of non-convex regularizations in this problem. Our contribution consists of three parts. First, we propose the leaky capped norm regularization (LCNR), which allows model weights below a certain threshold to be regularized more strongly as opposed to those above, therefore imposes strong sparsity and only introduces controllable estimation bias. We propose a majorization-minimization algorithm to optimize the joint objective function. Second, our study over monocular 3D shape recovery and neural networks with LCNR outperforms $\ell_1$ and other non-convex regularizations, achieving state-of-the-art performance and faster convergence. Third, we prove a theoretical global convergence speed on the 3D recovery problem. To the best of our knowledge, this is the first convergence analysis of the 3D recovery problem.

📄 PDF Abstract BibTeX arXiv:1711.02857

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

SPEED The monocular depth estimation (MDE) is the task of estimating depth from a single frame. This information is an essential knowledge in many computer vision tasks such as scene…

Similar Papers 제목 키워드 기반

Capped norm linear discriminant analysis and its applications

2020-11-04 · Jiakou Liu, Xiong Xiong, Pei-Wei Ren, Da Zhao 외

Classical linear discriminant analysis (LDA) is based on squared Frobenious norm and hence is sensitive to outliers and noise. To improve the robustness of LDA, in this paper, we introduce capped l_{2,1}-norm of a matrix…

Diffusion Leaky Zero Attracting Least Mean Square Algorithm and Its Performance Analysis

2018-10-29

Recently, the leaky diffusion least-mean-square (DLMS) algorithm has obtained much attention because of its good performance for high input eigenvalue spread and low signal-to-noise ratio (SNR). However, the leaky DLMS a…

Monocular 3D Pose Recovery via Nonconvex Sparsity with Theoretical Analysis

2018-12-29 · Jianqiao Wangni, Dahua Lin, Ji Liu, Kostas Daniilidis 외

For recovering 3D object poses from 2D images, a prevalent method is to pre-train an over-complete dictionary $\mathcal D=\{B_i\}_i^D$ of 3D basis poses. During testing, the detected 2D pose $Y$ is matched to dictionary …

Robust Capped lp-Norm Support Vector Ordinal Regression

2024-04-25 · Haorui Xiang, Zhichang Wu, Guoxu Li, Rong Wang 외

Ordinal regression is a specialized supervised problem where the labels show an inherent order. The order distinguishes it from normal multi-class problem. Support Vector Ordinal Regression, as an outstanding ordinal reg…

regression

Robust kernel-free quadratic surface twin support vector machine with capped $L_1$-norm distance metric

2024-05-27 · Qi Si, Zhi Xia Yang

Twin support vector machine (TSVM) is a very classical and practical classifier for pattern classification. However, the traditional TSVM has two limitations. Firstly, it uses the L_2-norm distance metric that leads to i…