---
title: "Learning Robust Multi-view Representation Using Dual-masked VAEs"
authors:
  - "Jiedong Wang"
  - "Kai Guo"
  - "Peng Hu"
  - "Xi Peng"
  - "Hao Wang"
doi: "10.24963/ijcai.2025/701"
url: "https://ijcai.org/proceedings/2025/701"
pdf: "https://ijcai.org/proceedings/2025/0701.pdf"
venue: "Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence"
year: 2025
pages: "6298-6306"
track: "Main Track"
keywords:
  - "Machine Learning: ML: Clustering"
  - "Machine Learning: ML: Representation learning"
  - "Machine Learning: ML: Multi-view learning"
---

# Learning Robust Multi-view Representation Using Dual-masked VAEs

Jiedong Wang, Kai Guo, Peng Hu, Xi Peng, Hao Wang

*Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence*, pages 6298-6306. DOI: [10.24963/ijcai.2025/701](https://doi.org/10.24963/ijcai.2025/701)

## Abstract

Most existing multi-view representation learning methods assume view-completeness and noise-free data. However, such assumptions are strong in real-world applications. Despite advances in methods tailored to view-missing or noise problems individually, a one-size-fits-all approach that concurrently addresses both remains unavailable. To this end, we propose a holistic method, called Dual-masked Variational Autoencoders (DualVAE), which aims at learning robust multi-view representation. The DualVAE exhibits an innovative amalgamation of dual-masked prediction, mixture-of-experts learning, representation disentangling, and a joint loss function in wrapping up all components. The key novelty lies in the dual-masked (view-mask and patch-mask) mechanism to mimic missing views and noisy data. Extensive experiments on four multi-view datasets show the effectiveness of the proposed method and its superior performance in comparison to baselines. The code is available at https://github.com/XLearning-SCU/2025-IJCAI-DualVAE.