AI Paper Atlas

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

Evidential Sparsification of Multimodal Latent Spaces in Conditional Variational Autoencoders

Masha Itkina, Boris Ivanovic, Ransalu Senanayake, Mykel J Kochenderfer, Marco Pavone

NeurIPS 2020 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Structured latent VAEs · 3 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 evidential sparse distribution instead of softmaxVQ-VAE PixelCNN prior · miniImageNet classification of samplesaccuracy (%): 20.688 -> 19.937No clear effect
use sparsemax instead of softmaxVQ-VAE PixelCNN prior · miniImageNet classification of samplesaccuracy (%): 20.688 -> 6.125Hurt
use evidential sparse distribution instead of sparsemaxVQ-VAE PixelCNN prior · miniImageNet classification of samplestop-5 accuracy (%): 17.500 -> 47.875Improved
use evidential sparse distribution instead of softmaxTrajectron++ · ETH pedestrian, direct samplingNLL: 4.698 -> 4.686No clear effect
use evidential sparse distribution instead of softmaxTrajectron++ · ETH pedestrian, top-5 samplingNLL: 3.951 -> 3.862Improved
use sparsemax instead of softmaxTrajectron++ · ETH pedestrian, top-5 samplingNLL: 3.951 -> 4.360Hurt
use sparsemax instead of softmaxTrajectron++ · ETH pedestrian, top-5 samplingFDE: 0.757 -> 0.802Hurt
use evidential sparse distribution instead of softmaxTrajectron++ · ETH pedestrian, top-5 samplingFDE: 0.757 -> 0.753No clear effect
use evidential sparse distribution instead of softmaxCVAE · MNISTBhattacharyya distance: 22%Improved
use evidential sparse distribution instead of sparsemaxCVAE · MNISTBhattacharyya distance: 10%Improved
use sparsemax instead of softmaxCVAE · MNIST odd querydistance to targetHurt
train on reduced fractions of MNIST (0.1, 0.5, 0.75, 1.0) (ablation)evidential sparse distribution · MNISTWasserstein and BhattacharyyaImproved

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