AI Paper Atlas

Research map / Image generation / GANs

LocalGAN: Modeling Local Distributions for Adversarial Response Generation

Baoxun Wang, Zhen Xu, Huan Zhang, Kexin Qiu, Deyuan Zhang, Chengjie Sun

JMLR 2021 · Image generation › GANs · approach: GANs for text generation · 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
use the DBM free-energy local distribution term instead of Euclidean distance ratio (ablation)LocalGAN · OpenSubtitlesDist-1: 0.0314 -> 0.036Improved
use the DBM free-energy local distribution term instead of Euclidean distance ratio (ablation)LocalGAN · OpenSubtitlesEnt4: 5.922 -> 6.073Improved
use the DBM free-energy local distribution term instead of Euclidean distance ratio (ablation)LocalGAN · WeiboEnt4: 8.134 -> 8.561Improved
use the DBM free-energy local distribution term instead of Euclidean distance ratio (ablation)LocalGAN · Weiborelevance: 0.311 -> 0.327Improved
use the DBM free-energy local distribution term instead of Euclidean distance ratio (ablation)LocalGAN · human evaluationinformativeness: 0.437 -> 0.536Improved
add the local-distribution hybrid loss to the GAN objective (beta 0 -> 0.9) (ablation)LocalGAN · OpenSubtitles and WeiboDist-1, Dist-2, Ent4, relevanceImproved
use only the local-distribution term (beta = 1.0) (ablation)LocalGAN · OpenSubtitles and WeiboDist-1, Dist-2, Ent4, relevanceHurt
add adversarial learning to Seq2Seq (Adver-REGS vs Seq2Seq)Seq2Seq · OpenSubtitlesEnt4: 5.650 -> 5.701Improved
add BigGAN operations (larger batch, doubled parameters, orthogonal regularization) to Adver-REGSAdver-REGS · WeiboEnt4: 7.658 -> 8.273Improved
replace baselines with LocalGAN for training stabilityadversarial NRG · OpenSubtitlesDist-1 std: 0.0004 -> 0.0006Improved
add LocalGAN over strongest baseline BigGANBigGAN · OpenSubtitlesrelevance: 1.117 -> 1.132Improved
add LocalGAN over strongest baseline BigGANBigGAN · human evaluationinformativeness: 0.487 -> 0.536Improved

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