AI Paper Atlas

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

Multimodal Generative Learning Utilizing Jensen-Shannon-Divergence

Thomas Sutter, Imant Daunhawer, Julia Vogt

NeurIPS 2020 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Score-based generative modeling analysis · 13 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
Use the mmJSD objective instead of MMVAE's objectivemultimodal VAE · MNIST-SVHN-Textlatent representation accuracy (Joint): 0.93 -> 0.98Improved
Use the mmJSD objective instead of MVAE's objectivemultimodal VAE · MNIST-SVHN-Textgenerated-sample classification accuracy (M from S,T): 0.22 -> 0.95Improved
Use the mmJSD objective instead of MMVAE's objectivemultimodal VAE · MNIST-SVHN-Textgenerated-sample classification accuracy (M from S,T): 0.91 -> 0.95Improved
Add modality-specific latent subspaces (ablation)mmJSD · MNIST-SVHN-Textgenerated-sample precision (M, S,T condition): 0.16 -> 0.64Improved
Add modality-specific latent subspaces (ablation)mmJSD · MNIST-SVHN-Textlatent representation accuracy (Joint): 0.98 -> 0.99Improved
Add modality-specific latent subspaces (ablation)multimodal VAE · MNIST-SVHN-Textgenerated-sample precision (M, S,T condition): mmJSD: 0.16 -> 0.64; MMVAE: 0.18 -> 0.63Mixed
Use mmJSD instead of MMVAEmultimodal VAE · CelebAlatent representation mean average precision (Joint): 0.42 -> 0.57Improved
Use mmJSD instead of MVAEmultimodal VAE · CelebAgeneration mean average precision (T to I): 0.30 -> 0.42Improved
Use mmJSD instead of MMVAEmultimodal VAE · CelebAgeneration mean average precision (T to I): 0.36 -> 0.42Improved
Use mmJSD instead of MVAEmultimodal VAE · MNIST-SVHN-Texttest set log-likelihood (joint model): -1864 -> -1961Hurt
Use modality-specific latent subspaces (ablation)MMVAE · MNIST-SVHN-Textgenerated-sample classification accuracy (S from M): 0.32 -> 0.88Improved
Use modality-specific latent subspaces (ablation)mmJSD · MNIST-SVHN-Textgenerated-sample classification accuracy (S from M): 0.37 -> 0.89Improved

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