AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

ShaLa: Multimodal Shared Latent Generative Modelling

Jiali Cui, Yan-Ying Chen, Yanxia Zhang, Matthew Klenk

AAAI 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Disentangled and discrete latent models · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
condition diffusion prior on embeddings h1:m instead of raw inputs x1:m (ablation)ShaLa · PolyMNIST unconditionalcoherence: 0.584 -> 0.815Improved
condition diffusion prior on embeddings h1:m instead of raw inputs x1:m (ablation)ShaLa · PolyMNIST conditionalcoherence: 0.612 -> 0.897Improved
condition diffusion prior on embeddings h1:m instead of raw inputs x1:m (ablation)ShaLa · MNIST-SVHN-Text unconditionalcoherence: 0.23 -> 0.44Improved
condition diffusion prior on embeddings h1:m instead of raw inputs x1:m (ablation)ShaLa · MNIST-SVHN-Text conditionalcoherence: 0.35 -> 0.75Improved
halve diffusion prior size (1/2x parameters) (ablation)ShaLa · PolyMNIST unconditionalFID: 47.30 -> 50.24Hurt
shrink diffusion prior to 1/4x parameters (ablation)ShaLa · PolyMNIST unconditionalFID: 47.30 -> 56.44Hurt
shrink diffusion prior to 1/4x parameters (ablation)ShaLa · PolyMNIST conditionalFID: 40.18 -> 49.75Hurt
double diffusion prior size (2x parameters) (ablation)ShaLa · PolyMNIST unconditionalFID: 47.30 -> 45.24Improved
double diffusion prior size (2x parameters) (ablation)ShaLa · PolyMNIST conditionalFID: 40.18 -> 37.15Improved
replace CMVAE with ShaLaCMVAE · PolyMNIST unconditionalcoherence: 0.781 -> 0.815Improved
replace CMVAE with ShaLaCMVAE · PolyMNIST conditionalcoherence: 0.897 -> 0.897No clear effect
replace CMVAE with ShaLaCMVAE · CUB conditionalFID: 28.00 -> 25.58Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy