DUAL: Dynamic Uncertainty-Aware Learning

2025/05/21 by Jiahao Qin, Bei Peng, Qin, Jiahao +7
Computer Science · #Advanced Neural Network Applications #Adversarial Robustness in Machine Learning #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2506.03158

openalex publication_date 2025/05/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Deep learning models frequently encounter feature uncertainty in diverse learning scenarios, significantly impacting their performance and reliability. This challenge is particularly complex in multi-modal scenarios, where models must integrate information from different sources with inherent uncertainties. We propose Dynamic Uncertainty-Aware Learning (DUAL), a unified framework that effectively handles feature uncertainty in both single-modal and multi-modal scenarios. DUAL introduces three key innovations: Dynamic Feature Uncertainty Modeling, which continuously refines uncertainty estimates through joint consideration of feature characteristics and learning dynamics; Adaptive Distribution-Aware Modulation, which maintains balanced feature distributions through dynamic sample influence adjustment; and Uncertainty-aware Cross-Modal Relationship Learning, which explicitly models uncertainties in cross-modal interactions. Through extensive experiments, we demonstrate DUAL's effectiveness across multiple domains: in computer vision tasks, it achieves substantial improvements of 7.1% accuracy on CIFAR-10, 6.5% accuracy on CIFAR-100, and 2.3% accuracy on Tiny-ImageNet; in multi-modal learning, it demonstrates consistent gains of 4.1% accuracy on CMU-MOSEI and 2.8% accuracy on CMU-MOSI for sentiment analysis, while achieving 1.4% accuracy improvements on MISR. The code will be available on GitHub soon.

Citations

• Detecting and Grounding Multi-Modal Media Manipulation and Beyond
• Robust Speech Recognition via Large-Scale Weak Supervision
• Feature Alignment by Uncertainty and Self-Training for Source-Free Unsupervised Domain Adaptation
• Multimodal Learning with Transformers: A Survey
• Representation Uncertainty in Self-Supervised Learning as Variational Inference
• Understanding and Improving Robustness of Vision Transformers through Patch-based Negative Augmentation
• A Survey of Uncertainty in Deep Neural Networks
• Improving robustness against common corruptions by covariate shift adaptation
• Evaluating Scalable Bayesian Deep Learning Methods for Robust Computer Vision
• Multi-attention Recurrent Network for Human Communication Comprehension
• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
• Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
• Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning

Related

• Uncertainty-Aware Multi-View Representation Learning
• MAP: Multimodal Uncertainty-Aware Vision-Language Pre-training Model
• Tailored Uncertainty Estimation for Deep Learning Systems
• Uncertainty-Resilient Multimodal Learning via Consistency-Guided Cross-Modal Transfer
• Uncertainty Baselines: Benchmarks for Uncertainty & Robustness in Deep Learning