Using Unsupervised Dynamic Feature Selection to Enhance Latent Representations

Bruno Corcuera Sánchez, Carlos Eiras-Franco, Brais Cancela

TL;DR: This paper presents a novel approach for enhancing latent representations using unsupervised Dynamic Feature Selection (DFS)

Abstract: Latent representations are critical for the performance and robustness of machine learning models, as they encode the essential features of data in a compact and informative manner. However, these representations are often affected by noisy or irrelevant features, which can degrade the model’s performance and generalization capabilities. This paper presents a novel approach for enhancing latent representations using unsupervised Dynamic Feature Selection (DFS). For each instance, the proposed method identifies and removes misleading or redundant information, ensuring that only the most relevant features contribute to the latent space. By leveraging an unsupervised framework, our approach avoids reliance on labeled data, making it broadly applicable across various domains and datasets. Experiments conducted on image datasets demonstrate that models equipped with unsupervised DFS achieve significant improvements in generalization performance across various tasks, including clustering and image generation, while maintaining a minimal increase in the computational cost.

Primary Area: General Machine Learning->Unsupervised and Semi-supervised Learning

Keywords: Dynamic Feature Selection, Latent Representations, Unsupervised learning, Self-supervised learning, Clustering, Image Generation, World Models

Submission Number: 6452