No more meta-parameter tuning in unsupervised sparse feature learning

2014/02/24 by Adriana Romero, Romero, Adriana, Petia Radeva +3 · 5 citations
Computer Science · Engineering · #Artificial intelligence #Blind Source Separation Techniques #Computer science #Discriminative model #Domain Adaptation and Few-Shot Learning #Engineering #Exploit #Face and Expression Recognition #Feature (linguistics) #Feature learning #Machine learning #Meta learning (computer science) #Pattern recognition (psychology) #Simple (philosophy) #Task (project management) #Unsupervised learning #cs.CV #cs.LG

paper · pdf · doi:10.48550/arxiv.1402.5766

published in arXiv (Cornell University) (Cornell University)

arxiv created 2014/02/24 · openalex publication_date 2014/02/24 · arxiv updated 2014/02/25 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/08

Abstract

We propose a meta-parameter free, off-the-shelf, simple and fast unsupervised feature learning algorithm, which exploits a new way of optimizing for sparsity. Experiments on STL-10 show that the method presents state-of-the-art performance and provides discriminative features that generalize well.

Citations

• Independent component analysis: algorithms and applications

Cited by

• CUNet: A Compact Unsupervised Network for Image Classification

Related

• Meta-Sparsity: Learning Optimal Sparse Structures in Multi-task Networks through Meta-learning
• Enhancing Unsupervised Feature Selection via Double Sparsity Constrained Optimization
• SLM: End-to-end Feature Selection via Sparse Learnable Masks
• Direct l_(2,p)-Norm Learning for Feature Selection
• Unsupervised Feature Analysis with Class Margin Optimization