Let the Prototype Guide You: Robust Aggregation of Sparse Multi-Class Annotations via Annotator Prototype Learning

Ju Chen, Jun Feng, Shenyu Zhang
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:14193-14207, 2026.

Abstract

Truth inference is a critical technique for aggregating noisy and biased multi-class classification annotations. State-of-the-art approaches model each annotator using an individual confusion matrix. While well-grounded, they suffer from two fundamental bottlenecks: 1) confusion matrices are underfit when annotators label only a small subset of tasks or when classes are imbalanced, and 2) a single confusion matrix per annotator is inadequate for capturing complex annotator behaviors, leading to class-level collapse when tasks are extremely difficult. Simultaneously addressing these challenges is non-trivial, as it demands both robustness to data sparsity and sufficient expressiveness for complex annotator patterns. In this paper, we propose CPBCC (Class-specific Prototype-driven Bayesian Classifier Combination), which creatively models annotators through a dual-pathway architecture: (i) learning class-specific prototype annotation patterns across all annotators, and (ii) learning annotator-specific weights over prototypes. This framework addresses the bottlenecks and achieves a robust yet rich annotator characterization. Experiments across 10 real-world datasets spanning five domains demonstrate that CPBCC yields a 26% accuracy improvement in the best case, and boosts average accuracy from 68.73% to 74.11%. Our source code is available at https://github.com/JuJuCHEN-HHU/CPBCC_PTBCC.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-chen26ac, title = {Let the Prototype Guide You: Robust Aggregation of Sparse Multi-Class Annotations via Annotator Prototype Learning}, author = {Chen, Ju and Feng, Jun and Zhang, Shenyu}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {14193--14207}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/chen26ac/chen26ac.pdf}, url = {https://proceedings.mlr.press/v306/chen26ac.html}, abstract = {Truth inference is a critical technique for aggregating noisy and biased multi-class classification annotations. State-of-the-art approaches model each annotator using an individual confusion matrix. While well-grounded, they suffer from two fundamental bottlenecks: 1) confusion matrices are underfit when annotators label only a small subset of tasks or when classes are imbalanced, and 2) a single confusion matrix per annotator is inadequate for capturing complex annotator behaviors, leading to class-level collapse when tasks are extremely difficult. Simultaneously addressing these challenges is non-trivial, as it demands both robustness to data sparsity and sufficient expressiveness for complex annotator patterns. In this paper, we propose CPBCC (Class-specific Prototype-driven Bayesian Classifier Combination), which creatively models annotators through a dual-pathway architecture: (i) learning class-specific prototype annotation patterns across all annotators, and (ii) learning annotator-specific weights over prototypes. This framework addresses the bottlenecks and achieves a robust yet rich annotator characterization. Experiments across 10 real-world datasets spanning five domains demonstrate that CPBCC yields a 26% accuracy improvement in the best case, and boosts average accuracy from 68.73% to 74.11%. Our source code is available at https://github.com/JuJuCHEN-HHU/CPBCC_PTBCC.} }
Endnote
%0 Conference Paper %T Let the Prototype Guide You: Robust Aggregation of Sparse Multi-Class Annotations via Annotator Prototype Learning %A Ju Chen %A Jun Feng %A Shenyu Zhang %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-chen26ac %I PMLR %P 14193--14207 %U https://proceedings.mlr.press/v306/chen26ac.html %V 306 %X Truth inference is a critical technique for aggregating noisy and biased multi-class classification annotations. State-of-the-art approaches model each annotator using an individual confusion matrix. While well-grounded, they suffer from two fundamental bottlenecks: 1) confusion matrices are underfit when annotators label only a small subset of tasks or when classes are imbalanced, and 2) a single confusion matrix per annotator is inadequate for capturing complex annotator behaviors, leading to class-level collapse when tasks are extremely difficult. Simultaneously addressing these challenges is non-trivial, as it demands both robustness to data sparsity and sufficient expressiveness for complex annotator patterns. In this paper, we propose CPBCC (Class-specific Prototype-driven Bayesian Classifier Combination), which creatively models annotators through a dual-pathway architecture: (i) learning class-specific prototype annotation patterns across all annotators, and (ii) learning annotator-specific weights over prototypes. This framework addresses the bottlenecks and achieves a robust yet rich annotator characterization. Experiments across 10 real-world datasets spanning five domains demonstrate that CPBCC yields a 26% accuracy improvement in the best case, and boosts average accuracy from 68.73% to 74.11%. Our source code is available at https://github.com/JuJuCHEN-HHU/CPBCC_PTBCC.
APA
Chen, J., Feng, J. & Zhang, S.. (2026). Let the Prototype Guide You: Robust Aggregation of Sparse Multi-Class Annotations via Annotator Prototype Learning. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:14193-14207 Available from https://proceedings.mlr.press/v306/chen26ac.html.

Related Material