AI Paper Atlas

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

Variational Mixture of HyperGenerators for Learning Distributions over Functions

Batuhan Koyuncu, Pablo Sanchez Martin, Ignacio Peis, Pablo M. Olmos, Isabel Valera

ICML 2023 · 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
Replace per-sample functa optimization with a single forward-pass encoder (VaMoH) for reconstructionreconstruction inference · PolyMNIST, ShapeS3D, CelebA-HQ, Era5, Chairsinference time (s): 2.29x to 3.64x faster than Functa (3 steps); 7.11x to 11.28x faster than Functa (10 steps)Improved
Use VaMoH instead of Functa for reconstructionimage-like datasets · PSNR test set, Era5 and ShapeS3DPSNRImproved
Use VaMoH instead of GASP for generationunconditional generation · CelebA-HQ 64x64FID: 66.27 vs 14.01 (GASP tr)Hurt
Use VaMoH instead of baselines for generationunconditional generation · ShapeS3D 64x64FID: 56.25 vs 57.81 (Functa) and 118.66 (GASP)Improved
Use VaMoH instead of baselines for generationunconditional generation · ShapeS3D 64x64recall: 0.64 vs 0.13 (Functa) and 0.16 (GASP)Improved
Use VaMoH instead of baselines for generationunconditional generation · CelebA-HQ 64x64recall: 0.0 vs 0.43 (GASP)Hurt
Flow-based (planar flow) prior replacing standard Gaussian prior on z (ablation)VaMoH latent prior · CelebA (qualitative)sample realism / reconstruction-generation alignmentImproved
Mixture with K=1 instead of K>1 hypergenerators (ablation)VaMoH · CelebA-HQ reconstructionreconstruction quality (qualitative)Hurt
Increase mixture components to K=10 (ablation)VaMoH · CelebA-HQ reconstructionreconstruction quality (qualitative)No clear effect
Point dropout during training (ablation)VaMoH encoder · CelebA-HQ patch imputationimputation quality (qualitative)Improved
Use Functa with normalizing-flow generator on PolyMNIST and CelebA-HQ instead of VaMoHFuncta baseline · PolyMNISTgeneration qualityHurt

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