Navigating Conflicting Views: Harnessing Trust for Learning
This work addresses the challenge of unreliable multi-view classification in real-world scenarios where views may not be perfectly aligned or equally important, offering a method to enhance model reliability, though it appears incremental as it builds on the Evidential Multi-view framework.
The paper tackled the problem of conflicting views in multi-view classification by developing a computational trust-based discounting method that accounts for instance-wise reliability of each view, resulting in improved prediction reliability as measured by metrics like Top-1 Accuracy and Multi-View Agreement with Ground Truth on six real-world datasets.
Resolving conflicts is critical for improving the reliability of multi-view classification. While prior work focuses on learning consistent and informative representations across views, it often assumes perfect alignment and equal importance of all views, an assumption rarely met in real-world scenarios, as some views may express distinct information. To address this, we develop a computational trust-based discounting method that enhances the Evidential Multi-view framework by accounting for the instance-wise reliability of each view through a probability-sensitive trust mechanism. We evaluate our method on six real-world datasets using Top-1 Accuracy, Fleiss' Kappa, and a new metric, Multi-View Agreement with Ground Truth, to assess prediction reliability. We also assess the effectiveness of uncertainty in indicating prediction correctness via AUROC. Additionally, we test the scalability of our method through end-to-end training on a large-scale dataset. The experimental results show that computational trust can effectively resolve conflicts, paving the way for more reliable multi-view classification models in real-world applications. Codes available at: https://github.com/OverfitFlow/Trust4Conflict