AI Paper Atlas

Research map / Benchmarks / ImageNet-A

ImageNet-A leaderboard

121 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on ImageNet-A. Best accuracy: 83.7 (Noisy Student Training (L2), CVPR 2020).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

ImageNet-A · accuracy

121 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
183.7Noisy Student Training (L2)
EfficientNet-L2
Self-Training With Noisy Student Improves ImageNet Classification
CVPR 2020 · Table: Robustness results on I
Knowledge distillation
281.92COTE
CLIP ViT-B/16
Controllable Federated Prompt Learning at Test Time
CVPR 2026 · Table 1
Personalized and heterogeneous FL
375.5ViT-g MVT +FT
EVA ViT-g
Machine Vision Therapy: Multimodal Large Language Models Can Enhance Visual Robustness via Denoising In-Context Learning
ICML 2024 · Table 1
Multimodal LLMs
466.59PAA
CLIP-ViT-B/16
Probe, Anchor, and Amend: Active Test-Time Adaptation of Vision-Language Models
ECCV 2026 · Table 1
Test-time adaptation
564.52SCAP
CLIP ViT-B/16
SCAP: Transductive Test-Time Adaptation via Supportive Clique-based Attribute Prompting
CVPR 2025 · Table 1
CLIP and zero-shot
664.38A2Memory
CLIP ViT-B/16
Learning to Memorize with Attributive and Associative Memory for Online Test-Time Adaptation of Vision-Language Models
ICML 2026 · Table 2
Test-time adaptation
764.29STS Ensemble
CLIP ViT-B/16
Test-Time Spectrum-Aware Latent Steering for Zero-Shot Generalization in Vision-Language Models
NeurIPS 2025 · Table: ImageNet and OOD varian
CLIP and zero-shot
863.6SATA-B*
ViT-Base/16
SATA: Spatial Autocorrelation Token Analysis for Enhancing the Robustness of Vision Transformers
CVPR 2025 · Table 1
Attention and transformers
963.09D2TPT
CLIP ViT-B/16
Doubly Debiased Test-Time Prompt Tuning for Vision-Language Models
AAAI 2026 · Table 2
CLIP and zero-shot
1061.15PTA
CLIP ViT-B/16
Prototype-Based Test-Time Adaptation of Vision-Language Models
ICML 2026 · Table 2
CLIP and zero-shot
111 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.