arXiv cs.AI· Alexandre Myara, Nicolas Bourriez, Thomas Boyer, Thomas Lemercier, Ihab Bendidi, Auguste Genovesio·· 1 天前AI 评分22
XFactors：用对比监督实现解耦信息瓶颈的弱监督 VAE 框架
XFACTORS: Disentangled Information Bottleneck via Contrastive Supervision
AI 导读
XFactors 是一个弱监督 VAE 框架，通过对比监督实现解耦信息瓶颈，将表示分解为残差子空间 S 与各因子专属子空间 T₁…T_K，用 InfoNCE 损失拉近同因子值隐变量、推开不匹配对，并以 KL 正则施加高斯结构，无需对抗训练和辅助分类器。
正文 · 原文
Abstract:Disentangled representation learning aims to map independent factors of variation to independent representation components. On one hand, purely unsupervised approaches have proven successful on fully disentangled synthetic data, but fail to recover semantic factors from real data without strong inductive biases. On the other hand, supervised approaches are unstable and hard to scale to large attribute sets because they rely on adversarial objectives or auxiliary classifiers.
We introduce \textsc{XFactors}, a weakly-supervised VAE framework that disentangles and provides explicit control over a chosen set of factors. Building on the Disentangled Information Bottleneck perspective, we decompose the representation into a residual subspace $\mathcal{S}$ and factor-specific subspaces $\mathcal{T}_1,\ldots,\mathcal{T}_K$ and a residual subspace $\mathcal{S}$. Each target factor is encoded in its assigned $\mathcal{T}_i$ through contrastive supervision: an InfoNCE loss pulls together latents sharing the same factor value and pushes apart mismatched pairs. In parallel, KL regularization imposes a Gaussian structure on both $\mathcal{S}$ and the aggregated factor subspaces, organizing the geometry without additional supervision for non-targeted factors and avoiding adversarial training and classifiers.
Across multiple datasets, with constant hyperparameters, \textsc{XFactors} achieves state-of-the-art disentanglement scores and yields consistent qualitative factor alignment in the corresponding subspaces, enabling controlled factor swapping via latent replacement. We further demonstrate that our method scales correctly with increasing latent capacity and evaluate it on the real-world dataset CelebA. Our code is available at \href{this https URL}{this http URL}.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2601.21688 [cs.LG] |
| (or arXiv:2601.21688v3 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2601.21688 arXiv-issued DOI via DataCite |
Submission history
From: Alexandre Myara [view email]
[v1]
Thu, 29 Jan 2026 13:20:48 UTC (44,758 KB)
[v2]
Thu, 9 Jul 2026 17:00:50 UTC (42,313 KB)
[v3]
Wed, 11 Nov 2026 14:32:31 UTC (46,990 KB)
来源：arXiv cs.AI · arxiv.org