Flexible disentangled representation learning with soft-splitting for multi-view data

Xunzhan Yao

doi:10.1016/j.imavis.2025.105722

Flexible disentangled representation learning with soft-splitting for multi-view data

2025-09-13

0

PRE

AI

DOI：10.1016/j.imavis.2025.105722

Original

Original request for help

Save

Abstract

En 中文

• To better disentangle the common and peculiar information within multi-view data, a novel adaptive soft-splitting multi-view gated fusion auto-encoder network is proposed. • We employ a soft-splitting mask to dynamically adjust the ratio between the common and peculiar information. • To better fuse the common embeddings, Sliced Wasserstein Distance is used to perform distribution alignment.

Flexible disentangled representation learning with soft-splitting for multi-view data

Abstract

Journal

Organization