AI Paper Atlas

Research map / Benchmarks / PACS

PACS leaderboard

64 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on PACS. Best accuracy: 97.63 (MoE2 (ours), AAAI 2026).

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.

PACS · accuracy

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

#accuracy ↑MethodPaperTopic
197.63MoE2 (ours)
CLIP ViT-B/16
MoE^2: A Mixture-of-Mixtures of Experts for Ensemble-Free Domain Generalization
AAAI 2026 · Table 1
Domain adaptation and generalization
296.7Single + Both
CLIP-ViT
Explore and Exploit the Diverse Knowledge in Model Zoo for Domain Generalization
ICML 2023 · Table 1
Domain adaptation and generalization
396.63OLOR-SGD
ConvNeXt-B
One Step Learning, One Step Review
AAAI 2024 · classification results table
CNN features
494.7CLIP L + CAD
CLIP L
Optimal Representations for Covariate Shift
ICLR 2022 · Table 1
Self-supervised and contrastive
593.8DVD
ResNet-50 CLIP-pretrained
Vicinity-Guided Discriminative Latent Diffusion for Privacy-Preserving Domain Adaptation
NeurIPS 2025 · Domain generalization table
Domain adaptation and generalization
693.48iMSDA
ResNet-18
Partial disentanglement for domain adaptation
ICML 2022 · Table 2
Domain adaptation and generalization
791.82Bayes-CAL
CLIP ResNet-50
Bayesian Cross-Modal Alignment Learning for Few-Shot Out-of-Distribution Generalization
AAAI 2023 · Table 2
Zero-/few-shot and meta-learning
891.41CLAdapter
ViT-B
Unleashing Foundation Vision Models: Adaptive Transfer for Diverse Data-Limited Scientific Domains
NeurIPS 2025 · Table, Results on Tiny-ImageNe
Recognition and long-tail
990.5Model ratatouille (greedy)
ResNet-50
Model Ratatouille: Recycling Diverse Models for Out-of-Distribution Generalization
ICML 2023 · Table 1
Recognition and long-tail
1090.3REMA
ResNet-50
Reconstruct and Match: Out-of-Distribution Robustness via Topological Homogeneity
NeurIPS 2024 · Table 3
Test-time adaptation
43 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

PACS · average accuracy

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

#average accuracy ↑MethodPaperTopic
196ProReg
CLIP ViT-B/32
Debiased Fine-Tuning for Vision-Language Models by Prompt Regularization
AAAI 2023 · Table 5
CLIP and zero-shot
292.6DGFamba
VMamba
DGFamba: Learning Flow Factorized State Space for Visual Domain Generalization
AAAI 2025 · PACS results table
Domain adaptation and generalization
389HYPO w/ SWAD
ResNet-50
HYPO: Hyperspherical Out-Of-Distribution Generalization
ICLR 2024 · Table 2
OOD and anomaly detection
488.4GENIE (all components)
ResNet-50
One-Step Generalization Ratio Guided Optimization for Domain Generalization
ICML 2025 · Table: Ablation study on the P
Optimizers
587.5DANN+ELS
ResNet50
Free Lunch for Domain Adversarial Training: Environment Label Smoothing
ICLR 2023 · VLCS and PACS results table
Noisy labels
686.9MODE-A
ResNet18
Moderately Distributional Exploration for Domain Generalization
ICML 2023 · Table 1
Domain adaptation and generalization
786.9Meta-DMoE
ResNet-50
Meta-DMoE: Adapting to Domain Shift by Meta-Distillation from Mixture-of-Experts
NeurIPS 2022 · Table 3
Domain adaptation and generalization
885.84DB + GSNR + meta
ResNet18
Domain Generalization Guided by Gradient Signal to Noise Ratio of Parameters
ICCV 2023 · Table 6
Domain adaptation and generalization
985.7COMEN
ResNet-18
Compound Domain Generalization via Meta-Knowledge Encoding
CVPR 2022 · Table 1
Domain adaptation and generalization
1084.63FedDAP
ResNet-18
FedDAP: Domain-Aware Prototype Learning for Federated Learning under Domain Shift
CVPR 2026 · Table 2
Personalized and heterogeneous FL
4 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.