ProPML: Probability Partial Multi-label Learning

Łukasz Struski, Adam Pardyl, Jacek Tabor, Bartosz Zieliński
3/12/2024
cs.LG

No official code found. Best implementation: gmum/propml (4 stars, last commit 2023-04-28).

Abstract

Partial Multi-label Learning (PML) is a type of weakly supervised learning where each training instance corresponds to a set of candidate labels, among which only some are true. In this paper, we introduce \our{}, a novel probabilistic approach to this problem that extends the binary cross entropy to the PML setup. In contrast to existing methods, it does not require suboptimal disambiguation and, as such, can be applied to any deep architecture. Furthermore, experiments conducted on artificial and real-world datasets indicate that \our{} outperforms existing approaches, especially for high noise in a candidate set.

View on arXivView PDF

Code Implementations(1)

gmum/propmlCommunity100%

Supplementary material for ProPML: Probability Partial Multi-label Learning

40PythonApr 28, 20233 years ago

Similar Papers

Deep Partial Multi-Label Learning with Graph Disambiguation

86% match2023

Exploiting Conjugate Label Information for Multi-Instance Partial-Label Learning

83% match1 repo92024

Adversary-Aware Partial label learning with Label distillation

82% match2023

CroSel: Cross Selection of Confident Pseudo Labels for Partial-Label Learning

82% match1 repo162023

Multi-Instance Partial-Label Learning with Margin Adjustment

82% match1 repo142026

Least-Ambiguous Multi-Label Classifier

82% match2025

Cite this paper

@article{struski2024propml,
  title  = {ProPML: Probability Partial Multi-label Learning},
  author = {Łukasz Struski and Adam Pardyl and Jacek Tabor and Bartosz Zieliński},
  year   = {2024},
  eprint = {2403.07603},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2403.07603v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2403.07603.svg)](https://codeofpaper.com/paper/2403.07603)

Discussion