AI Paper Atlas

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

Latent Stochastic Interpolants

Saurabh Singh, Dmitry Lagun

ICLR 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Hierarchical VAEs and posterior collapse · 1 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
train latent SI jointly with encoder and decoder (LSI) instead of observation-space SIStochastic interpolants · ImageNet 256x256FID: 3.87 -> 3.91No clear effect
train in latent space rather than pixel spaceStochastic interpolants · ImageNet 256x256GFLOPs per forward pass of generative model: 1288 -> 450Improved
joint training (beta > 0) instead of independent encoder-decoder (beta -> 0) (ablation)LSI · ImageNet 128x128FID: 4.53 -> 3.75Improved
increase beta beyond the best value (ablation)LSI · ImageNet 128x128FIDHurt
joint training when shifting capacity from latent model to encoder/decoder (k=6) (ablation)LSI · ImageNet 128x128FID: 4.87 -> 3.96Improved
use constant or learned encoder noise instead of a deterministic encoder (ablation)LSI · ImageNet 128x128FIDImproved
use the proposed parameterization instead of denoising, original flow or noise prediction (ablation)LSI · ImageNet 128x128 FID @1K epochsFID: 4.28 -> 3.76Improved
use Gaussian prior instead of Uniform, Laplacian or Gaussian Mixture p0 (ablation)LSI · ImageNet 128x128 FID @1K epochsFID: 4.26 (GMM) -> 3.76Improved
move first/last k=9 blocks from latent model into encoder/decoder (ablation)LSI · ImageNet 128x128FID: 3.76 -> 4.61Hurt
use fixed constant encoder noise instead of learned encoder variance (ablation)LSI · ImageNet 128x128FIDImproved

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