Saved in:
Bibliographic Details
Main Authors: Han, Yu, Ceross, Aaron, Bergmann, Jeroen H. M.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.00422
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912356993335296
author Han, Yu
Ceross, Aaron
Bergmann, Jeroen H. M.
author_facet Han, Yu
Ceross, Aaron
Bergmann, Jeroen H. M.
contents Accurate classification of medical device risk levels is essential for regulatory oversight and clinical safety. We present a Transformer-based multimodal framework that integrates textual descriptions and visual information to predict device regulatory classification. The model incorporates a cross-attention mechanism to capture intermodal dependencies and employs a self-training strategy for improved generalization under limited supervision. Experiments on a real-world regulatory dataset demonstrate that our approach achieves up to 90.4% accuracy and 97.9% AUROC, significantly outperforming text-only (77.2%) and image-only (54.8%) baselines. Compared to standard multimodal fusion, the self-training mechanism improved SVM performance by 3.3 percentage points in accuracy (from 87.1% to 90.4%) and 1.4 points in macro-F1, suggesting that pseudo-labeling can effectively enhance generalization under limited supervision. Ablation studies further confirm the complementary benefits of both cross-modal attention and self-training.
format Preprint
id arxiv_https___arxiv_org_abs_2505_00422
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Toward Automated Regulatory Decision-Making: Trustworthy Medical Device Risk Classification with Multimodal Transformers and Self-Training
Han, Yu
Ceross, Aaron
Bergmann, Jeroen H. M.
Machine Learning
Computation and Language
Accurate classification of medical device risk levels is essential for regulatory oversight and clinical safety. We present a Transformer-based multimodal framework that integrates textual descriptions and visual information to predict device regulatory classification. The model incorporates a cross-attention mechanism to capture intermodal dependencies and employs a self-training strategy for improved generalization under limited supervision. Experiments on a real-world regulatory dataset demonstrate that our approach achieves up to 90.4% accuracy and 97.9% AUROC, significantly outperforming text-only (77.2%) and image-only (54.8%) baselines. Compared to standard multimodal fusion, the self-training mechanism improved SVM performance by 3.3 percentage points in accuracy (from 87.1% to 90.4%) and 1.4 points in macro-F1, suggesting that pseudo-labeling can effectively enhance generalization under limited supervision. Ablation studies further confirm the complementary benefits of both cross-modal attention and self-training.
title Toward Automated Regulatory Decision-Making: Trustworthy Medical Device Risk Classification with Multimodal Transformers and Self-Training
topic Machine Learning
Computation and Language
url https://arxiv.org/abs/2505.00422