Research map / Label-efficient and robust learning / Recognition and long-tail

Complementary Label Learning with Positive Label Guessing and Negative Label Enhancement

ICLR 2025 · Label-efficient and robust learning › Recognition and long-tail · approach: Partial and multi-label learning

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Reformulates complementary label learning as an inverse problem and splits it into positive label guessing and negative label enhancement subtasks to recover full label information from complementary labels.

Problem

Complementary label learning (CLL) trains a multi-class classifier using only complementary labels, which specify classes an instance does not belong to.

Approach

Reformulates CLL as an inverse problem that infers full label information from output-space information, decomposed into positive label guessing and negative label enhancement subtasks.

Closest papers by meaning

• Learning with Complementary Labels Revisited: The Selected-Completely-at-Random Setting Is More Practical (ICML 2024 · similarity 0.91)
• Adaptive Integration of Partial Label Learning and Negative Learning for Enhanced Noisy Label Learning (AAAI 2024 · similarity 0.91)
• Addressing Multi-Label Learning with Partial Labels: From Sample Selection to Label Selection (AAAI 2025 · similarity 0.89)
• Partial Label Learning with a Partner (AAAI 2024 · similarity 0.89)
• Forming Auxiliary High-confident Instance-level Loss to Promote Learning from Label Proportions (CVPR 2025 · similarity 0.88)
• Realistic Evaluation of Deep Partial-Label Learning Algorithms (ICLR 2025 · similarity 0.88)