Partial multi-label learning aims to induce a multi-label predictor from inaccurately annotated examples, where a set of candidate labels is assigned to each training example but only some of thesecandidateonesarevalid.

Relational topic models: Derived from the traditional topic models [12, 29], RTMs [14, 15] are introduced to jointly consider the generations of document contents and their relationships.

However,existing methods lackrobustness as they may decide to explore areas of the latent space for which no data was available during training andwhere thedecoder canbeunreliable, leading tothe generation ofunrealistic orinvalidobjects. Wepropose toleveragetheepistemic uncertainty of the decoder to guide the optimization process.

In this work, we focus on the task of controlling the level of sparsity when performing sparse learning.

In this work, we focus on the task of controlling the level of sparsity when performing sparse learning.