<ul><li>Choosing a meaningful subset of features from high-dimensional observations in unsupervised settings can greatly enhance the accuracy of downstream analysis.</li><li>The proposed method is a self-supervised graph-based approach for unsupervised feature selection.</li><li>It involves computing robust pseudo-labels using the graph Laplacian's eigenvectors and a model stability criterion.</li><li>Experiments on real-world datasets demonstrate the method's effectiveness, especially in biological datasets.</li></ul>

Spectral Self-supervised Feature Selection

Discover more