1
How Neural Losses Shape VAE Latents
探索VAE隐空间如何被神经损失函数塑造，揭示潜变量表征的新视角
arXiv:2606.00635v1 Announce Type: new Abstract: Modern VAEs are rarely trained with the pointwise likelihood implied by the standard $\beta$-VAE objec…
2
Training-Free Vector Quantization via Gaussian VAEs
提出一种无需训练的向量量化新方法，利用高斯VAE实现高效表示学习，为量化领域开辟新路径。
arXiv:2512.06609v3 Announce Type: replace Abstract: Vector-quantized variational autoencoders (VQ-VAEs) are discrete autoencoders that compress images…
3
Commutator-Induced Uncertainty in VAEs
VAE新视角：用交换子数学工具量化模型不确定性，理论严谨有深度
arXiv:2605.23449v1 Announce Type: new Abstract: Variational autoencoders (VAEs) often struggle to represent non-commutative structure in learned laten…
4
Improved Baselines with Representation Autoencoders
用预训练视觉编码器替代传统 VAE，系统性设计选择研究揭示三大简化改进思路。
arXiv:2605.18324v1 Announce Type: cross Abstract: Representation Autoencoders (RAE) replace traditional VAE with pretrained vision encoders. In this p…
5
Isolating Nonlinear Independent Sources in fMRI with $\beta$-TCVAE Models
用β-TCVAE模型从脑功能磁共振数据中分离非线性独立源，揭示大脑网络隐藏信号。
arXiv:2605.16708v1 Announce Type: new Abstract: Learning meaningful latent representations from nonlinear fMRI data remains a fundamental challenge in…