Disentangled Representation Learning via Flow Matching
This work addresses a specific bottleneck in disentangled representation learning for generative modeling, offering incremental improvements.
The paper tackles the problem of semantic misalignment in disentangled representation learning by proposing a flow matching framework with an orthogonality regularizer to suppress cross-factor interference. Experiments show consistent improvements over baselines with higher disentanglement scores and better controllability and sample fidelity.
Disentangled representation learning aims to capture the underlying explanatory factors of observed data, enabling a principled understanding of the data-generating process. Recent advances in generative modeling have introduced new paradigms for learning such representations. However, existing diffusion-based methods encourage factor independence via inductive biases, yet frequently lack strong semantic alignment. In this work, we propose a flow matching-based framework for disentangled representation learning, which casts disentanglement as learning factor-conditioned flows in a compact latent space. To enforce explicit semantic alignment, we introduce a non-overlap (orthogonality) regularizer that suppresses cross-factor interference and reduces information leakage between factors. Extensive experiments across multiple datasets demonstrate consistent improvements over representative baselines, yielding higher disentanglement scores as well as improved controllability and sample fidelity.