Abstract: Partial-label learning (PLL) is an important weakly supervised learning problem, which allows each training example to have a candidate label set instead of a single ground-truth label.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results