Saved in:
Bibliographic Details
Main Authors: Zhang, Liyun, Lian, Zheng, Liu, Hong, Takebe, Takanori, Nakashima, Yuta
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.09525
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916885212168192
author Zhang, Liyun
Lian, Zheng
Liu, Hong
Takebe, Takanori
Nakashima, Yuta
author_facet Zhang, Liyun
Lian, Zheng
Liu, Hong
Takebe, Takanori
Nakashima, Yuta
contents Multi-annotator learning (MAL) aims to model annotator-specific labeling patterns. However, existing methods face a critical challenge: they simply skip updating annotator-specific model parameters when encountering missing labels, i.e., a common scenario in real-world crowdsourced datasets where each annotator labels only small subsets of samples. This leads to inefficient data utilization and overfitting risks. To this end, we propose a novel similarity-weighted semi-supervised learning framework (SimLabel) that leverages inter-annotator similarities to generate weighted soft labels for missing annotations, enabling the utilization of unannotated samples rather than skipping them entirely. We further introduce a confidence-based iterative refinement mechanism that combines maximum probability with entropy-based uncertainty to prioritize predicted high-quality pseudo-labels to impute missing labels, jointly enhancing similarity estimation and model performance over time. For evaluation, we contribute a new multimodal multi-annotator dataset, AMER2, with high and more variable missing rates, reflecting real-world annotation sparsity and enabling evaluation across different sparsity levels.
format Preprint
id arxiv_https___arxiv_org_abs_2504_09525
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SimLabel: Similarity-Weighted Iterative Framework for Multi-annotator Learning with Missing Annotations
Zhang, Liyun
Lian, Zheng
Liu, Hong
Takebe, Takanori
Nakashima, Yuta
Multimedia
Artificial Intelligence
Multi-annotator learning (MAL) aims to model annotator-specific labeling patterns. However, existing methods face a critical challenge: they simply skip updating annotator-specific model parameters when encountering missing labels, i.e., a common scenario in real-world crowdsourced datasets where each annotator labels only small subsets of samples. This leads to inefficient data utilization and overfitting risks. To this end, we propose a novel similarity-weighted semi-supervised learning framework (SimLabel) that leverages inter-annotator similarities to generate weighted soft labels for missing annotations, enabling the utilization of unannotated samples rather than skipping them entirely. We further introduce a confidence-based iterative refinement mechanism that combines maximum probability with entropy-based uncertainty to prioritize predicted high-quality pseudo-labels to impute missing labels, jointly enhancing similarity estimation and model performance over time. For evaluation, we contribute a new multimodal multi-annotator dataset, AMER2, with high and more variable missing rates, reflecting real-world annotation sparsity and enabling evaluation across different sparsity levels.
title SimLabel: Similarity-Weighted Iterative Framework for Multi-annotator Learning with Missing Annotations
topic Multimedia
Artificial Intelligence
url https://arxiv.org/abs/2504.09525