Robust Label Proportions Learning

Yuheng Jia (Southeast University) · Jueyu Chen (Southeast University) · Wantao Wen (Southeast University) · Yeqiang Wang (Northwest A&F University) · Erliang Lin (Southeast University) · Yemin Wang (Xiamen University)

ablation studiesauxiliary classifierbag-level label proportionsencoder pretraininghigh-confidence setsinstance-level classifierslearning from label proportionsllp-otd mechanismllpmixlow-confidence setspseudo-label noisepseudo-labeling state-of-the-art performancetraining refinementunsupervised contrastive learningweakly-supervised learning

Learning from Label Proportions (LLP) is a weakly-supervised paradigm that uses bag-level label proportions to train instance-level classifiers, offering a practical alternative to costly instance-level annotation. However, the weak supervision makes effective training challenging, and existing methods often rely on pseudo-labeling, which introduces noise. To address this, we propose RLPL, a two-stage framework. In the first stage, we use unsupervised contrastive learning to pretrain the encoder and train an auxiliary classifier with bag-level supervision. In the second stage, we introduce an LLP-OTD mechanism to refine pseudo labels and split them into high- and low-confidence sets. These sets are then used in LLPMix to train the final classifier. Extensive experiments and ablation studies on multiple benchmarks demonstrate that RLPL achieves comparable state-of-the-art performance and effectively mitigates pseudo-label noise.