AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

ArcVQ-VAE: A Spherical Vector Quantization Framework with ArcCosine Additive Margin

Jaeyung Kim, YoungJoon Yoo

ICML 2026 · Image generation › Autoregressive and latent generation · approach: Hierarchical VQ-VAE 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
add Ball-Bounded Norm Regularization to VQ-VAE (ablation)VQ-VAE · MNISTrFID: 3.43 -> 2.71Improved
add Ball-Bounded Norm Regularization to VQ-VAE (ablation)VQ-VAE · CIFAR10rFID: 39.67 -> 35.18Improved
add ArcCosine Additive Margin Loss on top of BBNR (ablation)VQ-VAE + BBNR · MNISTrFID: 2.71 -> 1.68Improved
add ArcCosine Additive Margin Loss on top of BBNR (ablation)VQ-VAE + BBNR · CIFAR10rFID: 35.18 -> 26.91Improved
vary top-k of nearest latent tokens (1, 3, 8) (ablation)ArcVQ-VAE · MNIST and CIFAR10PSNR / rFID: CIFAR10 rFID 27.08, 26.91, 27.02No clear effect
increase additive angular margin m from 0.1 to 1.0 (ablation)ArcVQ-VAE · CIFAR10rFID: 26.91 -> 28.52Hurt
increase scaling factor s from 10 to 20 (ablation)ArcVQ-VAE · CIFAR10PSNR / rFID: PSNR 24.78 -> 23.95; rFID 26.91 -> 28.90Hurt
increase codebook dimension from 4 to 16 (ablation)ArcVQ-VAE · FFHQcodebook usage vs rFID: usage 98% -> 78%; rFID 2.71 -> 2.14Mixed
apply SAMP (BBNR + ArcLoss) to vanilla VQGAN tokenizerVQGAN · ImageNet-1K, 16x16 tokens, K=1024rFID: 7.94 -> 1.63Improved
use ArcVQ-VAE instead of CVQ-VAE at 16x16 tokensImageNet-1K tokenizer · ImageNet-1K, 16x16 tokensrFID: 1.57 -> 1.63Hurt
use ArcVQ-VAE instead of CVQ-VAE on CIFAR10small-scale tokenizer · CIFAR10rFID: 24.73 -> 26.91Hurt
use ArcVQ-VAE tokenizer instead of CVQ-VAE for LDM generationLDM on ImageNet-1K · ImageNet-1K generationFID: 6.87 -> 6.79Improved

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