On the prediction loss of the lasso in the partially labeled setting
This work addresses risk estimation in semi-supervised learning for researchers, but it is incremental as it builds on existing lasso methods with new adaptations and bounds.
The paper tackles the problem of predicting risk bounds for the lasso estimator in partially labeled regression settings, establishing non-asymptotic upper bounds that show how unlabeled features can positively impact prediction when design matrix properties are poor.
In this paper we revisit the risk bounds of the lasso estimator in the context of transductive and semi-supervised learning. In other terms, the setting under consideration is that of regression with random design under partial labeling. The main goal is to obtain user-friendly bounds on the off-sample prediction risk. To this end, the simple setting of bounded response variable and bounded (high-dimensional) covariates is considered. We propose some new adaptations of the lasso to these settings and establish oracle inequalities both in expectation and in deviation. These results provide non-asymptotic upper bounds on the risk that highlight the interplay between the bias due to the mis-specification of the linear model, the bias due to the approximate sparsity and the variance. They also demonstrate that the presence of a large number of unlabeled features may have significant positive impact in the situations where the restricted eigenvalue of the design matrix vanishes or is very small.