A curated list of awesome online test-time adaptation resources. Your contributions are always welcome!
-
T3A[Iwasawa and Matsuo, Proc. NeurIPS 2021] Test-time classifier adjustment module for model-agnostic domain generalization [PDF] [G-Scholar] [CODE] -
PAD[Wu et al., Proc. NeurIPS Workshops 2021] Domain-agnostic test-time adaptation by prototypical training with auxiliary data [PDF] [G-Scholar] -
NOTE[Gong et al., Proc. NeurIPS 2022] NOTE: Robust continual test-time adaptation against temporal correlation [PDF] [G-Scholar] [CODE] -
CoTTA[Wang et al., Proc. CVPR 2022] Continual test-time domain adaptation [PDF] [G-Scholar] [CODE] -
SAR[Niu et al., Proc. ICLR 2023] Towards stable test-time adaptation in dynamic wild world [PDF] [G-Scholar] [CODE] -
FEDTHE+[Jiang and Lin, Proc. ICLR 2023] Test-time robust personalization for federated learning [PDF] [G-Scholar] [CODE] -
...[Wu et al., Proc. CVPR 2023] Learning to adapt to online streams with distribution shifts [PDF] [G-Scholar--] -
VDP[Gan et al., Proc. AAAI 2023] Decorate the newcomers: Visual domain prompt for continual test time adaptation [PDF] [G-Scholar]
-
OSVOS[Voigtlaender and Leibe., Proc. BMVC 2017] Online adaptation of convolutional neural networks for video object segmentation [PDF] [G-Scholar] -
JITNet[Mullapudi et al., Proc. ICCV 2019] Online model distillation for efficient video inference [PDF] [G-Scholar] [CODE] -
MLDG+SIB[Zhang et al., Pattern Recognition 2022] Generalizable model-agnostic semantic segmentation via target-specific normalization [PDF] [G-Scholar] -
OASIS[Volpi et al., Proc. CVPR 2022] On the road to online adaptation for semantic image segmentation [PDF] [G-Scholar] [CODE] -
FTEA[Zhang et al., Arxiv 2022] Unseen object instance segmentation with fully test-time RGB-D embeddings adaptation [PDF] [G-Scholar] -
TransAdapt[Das et al., Proc. ICASSP 2023] TransAdapt: A transformative framework for online test time adaptive semantic segmentation [PDF] [G-Scholar--]
-
ViTTA[Lin et al., Arxiv 2022] Video test-time adaptation for action recognition [PDF] [G-Scholar] [CODE--] -
AUTO[Yang et al., Arxiv 2023] AUTO: Adaptive outlier optimization for test-time OOD detection [PDF--] [G-Scholar--] -
BOA[Guan et al., Proc. CVPR 2021] Bilevel online adaptation for out-of-domain human mesh reconstruction [PDF] [G-Scholar] [CODE] -
...[Kundu et al., Proc. CVPR 2022] Uncertainty-aware adaptation for self-supervised 3D human pose estimation [PDF] [G-Scholar]
-
ONDA[Mancini et al., Proc. IROS 2018] Kitting in the wild through online domain adaptation [PDF] [G-Scholar] -
Tent[Sun et al., Proc. ICLR 2021] Tent: Fully test-time adaptation by entropy minimization [PDF] [G-Scholar] [CODE] -
BACS[Zhou and Levine, Proc. NeurIPS 2021] Bayesian adaptation for covariate shift [PDF] [G-Scholar] -
TTA-PR[Sivaprasad and Fleuret, Proc. NeurIPS Workshops 2021] Test time adaptation through perturbation robustness [PDF] [G-Scholar] -
Core (alpha-BN)[You et al., Arxiv 2021] Test-time batch statistics calibration for covariate shift [PDF] [G-Scholar] -
SLR+IT[Mummadi et al., Arxiv 2021] Test-time adaptation to distribution shift by confidence maximization and input transformation [PDF] [G-Scholar] -
MixNorm[Hu et al., Arxiv 2021] MixNorm: Test-time adaptation through online normalization estimation [PDF] [G-Scholar] -
DLTTA[Yang et al., IEEE TMI 2022] DLTTA: Dynamic learning rate for test-time adaptation on cross-domain medical images [PDF] [G-Scholar] [CODE] -
EATA[Niu et al., Proc. ICML 2022] Efficient test-time model adaptation without forgetting [PDF] [G-Scholar] [CODE] -
VMP[Jing et al., Proc. NeurIPS 2022] Variational model perturbation for source-free domain adaptation [PDF] [G-Scholar] [CODE] -
TTAC[Su et al., Proc. NeurIPS 2022] Revisiting realistic test-time training: Sequential inference and adaptation by anchored clustering [PDF] [G-Scholar] [CODE] -
Conjugate PL[Goyal et al., Proc. NeurIPS 2022] Test-time adaptation via conjugate pseudo-labels [PDF] [G-Scholar] [CODE] -
DUA[Mirza et al., Proc. CVPR 2022] The norm must go on: Dynamic unsupervised domain adaptation by normalization [PDF] [G-Scholar] [CODE] -
LAME[Boudiaf et al., Proc. CVPR 2022] Parameter-free online test-time adaptation [PDF] [G-Scholar] [CODE] -
SWR-NSP[Choi et al., Proc. ECCV 2022] Improving test-time adaptation via shift-agnostic weight regularization and nearest source prototypes [PDF] [G-Scholar] -
CFA[Kojima et al., Proc. IJCAI 2022] Robustifying vision transformer without retraining from scratch by test-time class-conditional feature alignment [PDF] [G-Scholar] [CODE] -
MuSLA[Kingetsu et al., Proc. ICIP 2022] Multi-step test-time adaptation with entropy minimization and pseudo-labeling [PDF] [G-Scholar--] -
...[Bhardwaj et al., Proc. ISPASS 2022] Benchmarking test-time unsupervised deep neural network adaptation on edge devices [PDF] [G-Scholar] -
...[Bhardwaj et al., Proc. DATE 2022] Unsupervised test-time adaptation of deep neural networks at the edge: a case study [PDF] [G-Scholar] -
...[Kerssies et al., Arxiv 2022] Evaluating continual test-time adaptation for contextual and semantic domain shifts [PDF] [G-Scholar] -
ActMAD[Mirza et al., Arxiv 2022] ActMAD: Activation matching to align distributions for test-time-training [PDF] [G-Scholar--] -
GpreBN[Yang et al., Arxiv 2022] Test-time batch normalization [PDF] [G-Scholar] -
CAFA[Jung et al., Arxiv 2022] CAFA: Class-aware feature alignment for test-time adaptation [PDF] [G-Scholar] -
CAFe[Adachi et al., Arxiv 2022] Covariance-aware feature alignment with pre-computed source statistics for test-time adaptation [PDF] [G-Scholar] -
RMT[Döbler et al., Arxiv 2022] Robust mean teacher for continual and gradual test-time adaptation [PDF] [G-Scholar] [CODE] -
AdaODM[Zhang and Chen, Arxiv 2022] Adaptive domain generalization via online disagreement minimization [PDF] [G-Scholar] -
TAST[Jang and Chung, Proc. ICLR 2023] Test-time adaptation via self-training with nearest neighbor information [PDF] [G-Scholar] -
...[Wang and Wibisono, Proc. ICLR 2023] Towards understanding GD with hard and conjugate pseudo-labels for test-time adaptation [PDF] [G-Scholar] -
MECTA[Hong et al., Proc. ICLR 2023] MECTA: Memory-economic continual test-time model adaptation [PDF] [G-Scholar--] -
DELTA[Zhao et al., Proc. ICLR 2023] DELTA: Degradation-free fully test-time adaptation [PDF] [G-Scholar] -
PETAL[Brahma and Rai, Proc. CVPR 2023] A probabilistic framework for lifelong test-time adaptation [PDF] [G-Scholar] -
NHL[Tang et al., Proc. CVPR 2023] Neuro-modulated hebbian learning for fully test-time adaptation [PDF] [G-Scholar--] -
EcoTTA[Song et al., Proc. CVPR 2023] EcoTTA: Memory-efficient continual test-time adaptation via self-distilled regularization [PDF] [G-Scholar] -
TIPI[Nguyen et al., Proc. CVPR 2023] TIPI: Test time adaptation with transformation invariance [PDF--] [G-Scholar--] -
...[Yuan et al., Proc. CVPR 2023] Robust test-time adaptation in dynamic scenarios [PDF--] [G-Scholar--] -
...[Li et al., Proc. ICRA 2023] Test-time domain adaptation for monocular depth estimation [PDF--] [G-Scholar--] -
TeSLA[Tomar et al., Proc. CVPR 2023] TeSLA: Test-time self-learning With automatic adversarial augmentation [PDF--] [G-Scholar--] -
ECL[Han et al., Arxiv 2023] Rethinking precision of pseudo label: Test-time adaptation via complementary learning [PDF] [G-Scholar--] -
VPL[Ambekar et al., Misc 2023] Variational pseudo labels for meta test-time adaptation [PDF] [G-Scholar--]
ETLT[Fan et al., Arxiv 2022] A simple test-time method for out-of-distribution detection [PDF] [G-Scholar]
Dent[Wang et al., Arxiv 2021] Fighting gradients with gradients: Dynamic defenses against adversarial attacks [PDF] [G-Scholar] [CODE]
-
RNCR[Hu et al., Proc. MICCAI 2021] Fully test-time adaptation for image segmentation [PDF] [G-Scholar] -
...[Kuznietsov et al., Proc. WACV Workshops 2022] Towards unsupervised online domain adaptation for semantic segmentation [PDF] [G-Scholar] -
MM-TTA[Shin et al., Proc. CVPR 2022] MM-TTA: Multi-Modal test-time adaptation for 3D semantic segmentation [PDF] [G-Scholar] [CODE--] -
CD-TTA[Song et al., Arxiv 2022] CD-TTA: Compound domain test-time adaptation for semantic segmentation [PDF] [G-Scholar]
-
...[Peirone, Thesis 2022] EGO-T3: Test time training for egocentric videos [PDF] [G-Scholar] -
TeCo[Yi et al., Proc. ICLR 2023] Temporal coherent test-time optimization for robust video classification [PDF] [G-Scholar--]
GAPGC[Chen et al., Proc. ICML Workshops 2022] GraphTTA: Test time adaptation on graph neural networks [PDF] [G-Scholar]
-
Ev-TTA[Kim et al., Proc. CVPR 2022] Ev-TTA: Test-time adaptation for event-based object recognition [PDF] [G-Scholar] [CODE] -
OIL[Ye et al., Proc. EMNLP Findings 2022] Robust question answering against distribution shifts with test-time adaptation: An empirical study [PDF] [G-Scholar--] [CODE] -
MEMO-CL[Singh and Ortega, Proc. AAAI Workshops 2023] Addressing distribution shift at test time in pre-trained language models [PDF] [G-Scholar] -
ODR[Park and D'Amico, Arxiv 2022] Robust multi-task learning and online refinement for spacecraft pose estimation across domain gap [PDF] [G-Scholar] [CODE] -
OAP[Belli et al., Proc. ICIP 2022] Online adaptive personalization for face anti-spoofing [PDF] [G-Scholar] -
OTF[Lumentut and Park, Proc. ACMMM 2022] 3D body reconstruction revisited: Exploring the test-time 3D body mesh refinement strategy via surrogate adaptation [PDF] [G-Scholar--] -
TTA-COPE[Lee et al., Proc. CVPR 2023] TTA-COPE: Test-time adaptation for category-level object pose estimation [PDF--] [G-Scholar--]