Test-Time Augmentation

Follow this topic Ask about this topic

Test-time augmentation studies whether transforming inputs at inference and combining predictions can improve accuracy, robustness, uncertainty, or calibration without retraining. Methods include voting and ensembling across flips, scales, translations, or semantic rephrasings; inverse-aligning dense outputs; filtering views by consistency; and, in some settings, adapting prompts or model parameters. Evaluations span image classification and segmentation, distribution shifts and adversarial settings, vision-language spatial reasoning, and language-model reasoning, often comparing gains with single-pass inference at matched compute. Results are task-dependent: transformations must preserve relevant semantics and suit the model, since distribution shifts and added latency can erase gains, leaving reliable view selection and aggregation as key challenges.

Latest papers

20 papers1 paper in the last four weeks, down 75% on the four weeks before. 0.0% of all new papers.

Counterfactual Test-Time Augmentation for Shortcut Removal in Language Models
Oct 24, 2026Katie Matton, Rosalind W. Picard, John Guttag +1Shortcut Learning in Language Models Test-Time Augmentation
MIT · University of Michigan
Improving Calibration of Black-Box Radiology AI Using Test-Time Augmentation
Sep 24, 2026Nathan Le, Magdalini Paschali, Arogya Koirala +5Post-Hoc Calibration 3D Medical Imaging
Department of Radiology, Stanford University, Stanford, CA, USA · Department of Biomedical Data Science, Stanford University, Stanford, CA, USA · Department of Anesthesia, Intensive Care Medicine, and Pain Medicine, Medical University of Vienna, Spitalgasse 23, 1090 Vienna, Austria
INTCORT: Training-Free Spatial Reasoning Enhancement for Vision-Language Models via Input Transformations and Confidence Routing
Sep 21, 2026Haoran Sun, Jingqi Xu, Yanhui Li +3Visual Spatial Reasoning VLM Reasoning
The University of Hong Kong · University of Southern California · China Telecom +3
When Online Adaptation Hurts: Parameter-Frozen Test-Time Ensembling for Continual Medical Image Segmentation
Sep 18, 2026Ruijie HuangMedical Image Segmentation Continual Test-Time Adaptation
Hong Kong Baptist University, Faculty of Science and Technology
Branch-Centric Tokenization and Test-Time Augmentation for Skeleton Generation
Sep 5, 2026Zhengyuan Li, Chuanyu Pan, Yuanming Hu +1Test-Time Augmentation 3D Human Reconstruction
Purdue University · Meshy AI
GATTA: Graph Active Learning with Test-Time Augmentation
Aug 15, 2026Zsombor Bánfi, András Gézsi, András FormanekGraph Data Augmentation Active Learning
Department of Artificial Intelligence and Systems Engineering, Budapest University of Technology and Economics, Magyar tudósok krt. 2, Budapest, H-1117, Hungary. · Department of Electrical Engineering (ESAT), STADIUS Center for Dynamical Systems, Signal Processing and Data Analytics, KU Leuven, 3001 Leuven, Belgium.
PPOM: Marginalizing Patch-Grid Phase for CLIP-Based Generalizable Vision-Language Prompt Tuning
Aug 14, 2026Liang Wang, Haoyang Li, Chao Wang +3VLM Adaptation Prompt Learning for VLMs
Shanghai University, Shanghai, China · University of Technology Sydney, Sydney, Australia
Test-Time Augmentation for LLMs: Where to Inject Diversity at Inference Time
Aug 10, 2026Nikita Kozodoi, Zainab Afolabi, Jack ButlerLLM Inference Efficiency Self-Consistency Decoding
Amazon Web Services
Test-Time Augmentation for Tabular-to-Image Classifiers under Distribution Shifts
Aug 4, 2026Malena Loza, Felipe Grijalva, Eva Milara +3Distribution Shift Robustness Distribution Shift
Colegio de Ciencias e Ingenierías, Universidad San Francisco de Quito (USFQ), Quito, Ecuador · Department of Signal Theory and Communications, Telematics and Computing Systems, Rey Juan Carlos University, Madrid, Spain · Faculty of Experimental Sciences, Universidad Francisco de Vitoria, Madrid, Spain
Music-Source-Separation-Training (MSST): A Unified Framework for Training and Evaluating Music Demixing Models
Jul 25, 2026Roman Solovyev, Ilya Kiselev, Alexander Stempkovskiy +1Audio Source Separation Test-Time Augmentation
National Research University Higher School of Economics (HSE University), MIEM HSE · AlphaChip LLC. · Independent Researcher
Anatomy of a Sound Neural Reasoner: One-Shot Amortization, First-Pass Poisoning, and Search Inertness in Clue-Rich Completion
Jul 22, 2026Aleksey KomissarovSudoku Solving Inference-Time Search
Neapolis University, Pafos, Cyprus
USE: A Unified Self-Ensembling Framework for Test-Time Prompt Tuning
Jul 4, 2026Siru Jiang, Jian Liang, Ran He +1Ensemble Learning VLM Adaptation
School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences · NLPR & MAIS, Institute of Automation, Chinese Academy of Sciences · School of Artificial Intelligence, University of Chinese Academy of Sciences +1
Reliability-Guided Adaptive Ensembling for Robust Test-Time Adaptation
Jun 21, 2026Adam Koziak, Yuhong GuoTest-Time Adaptation Ensemble Learning
School of Computer Science, Carleton University, Ottawa, Canada · Canada CIFAR AI Chair, Amii, Canada
SAM3 Self-Distillation for Fine-Grained GOOSE 2D Semantic Segmentation
Jun 18, 2026Xuesong WangVision Foundation Models Self-Distillation
Wayne State University, Detroit, MI, USA
Phase Marginalization for Patch-Grid Instability in Vision Transformers
Jun 6, 2026Oğuzhan ErcanVision Transformer Dense Prediction
Scientific and Technological Research Council of Türkiye
What are the Right Symmetries for Formal Theorem Proving?
May 21, 2026Krzysztof Olejniczak, Radoslav Dimitrov, Xingyue Huang +3Large Language Model-Guided Theorem Proving Automated Theorem Proving
University of Oxford · KAIST · TU Wien +1
SegTTA: Training-Free Test-Time Augmentation for Zero-Shot Medical Imaging Segmentation
Apr 19, 2026Yihong Yao, Chunlei Li, Canxuan Gang +4Image Segmentation Medical Imaging Foundation Models
AI Geeks · Qingdao Municipal Hospital · University of Chinese Academy of Sciences
TTVS: Boosting Self-Exploring Reinforcement Learning via Test-time Variational Synthesis
Apr 9, 2026Sikai Bai, Haoxi Li, Jie Zhang +2Test-Time Adaptation Test-Time RL
The Hong Kong University of Science and Technology
I Can't Believe TTA Is Not Better: When Test-Time Augmentation Hurts Medical Image Classification
Apr 6, 2026Daniel Nobrega MedeirosTest-Time Augmentation Data Augmentation
University of Colorado at Boulder, MSc in Artificial Intelligence.
Invariant Transformation and Resampling based Epistemic-Uncertainty Reduction
Feb 26, 2026Sha HuInference-Time Optimization Test-Time Augmentation
Lund Research Center, Huawei Technologies Sweden AB, Sweden.

Oct 24, 2026J/K move · Enter open · S save

Counterfactual Test-Time Augmentation for Shortcut Removal in Language Models

Katie Matton, Rosalind W. Picard, John Guttag, Maggie Makar

MIT · University of Michigan

Shortcut Learning in Language Models Test-Time Augmentation Language Model Robustness Inference-Time Adaptation of Language Models

Open paper AskPDF arXiv

Large language models (LLMs) can exploit shortcuts, i.e., correlations present in training data that may not hold at test time. This can lead to poor predictions when correlations shift. Most existing solutions for removing shortcuts operate at training time, requiring labeled training data, model weight access, and significant computational resources. Here, we study the problem of test-time shortcut removal. We propose a family of counterfactual test-time augmentation (CTTA) methods: given an example, each method uses an auxiliary model to generate counterfactual edits in which the shortcut value is changed, and then takes a weighted average of the LLM's predictions across these edits. We prove that the optimal shortcut-invariant predictor is a member of this CTTA family, and we characterize its weights. The weights depend on the training distribution, and therefore are, in general, not identifiable from the trained model and unlabeled test data. However, we present assumptions under which the weights can be obtained using the token probabilities of the trained LLM, and propose tests to detect violations of these assumptions. We use these tests to develop Select-CTTA, an aggregate method that selects among CTTA methods. Empirically, we find that Select-CTTA outperforms test-time baselines and approaches the performance of models fine-tuned on shortcut-balanced data.