Accuracy on in-distribution data correlates with out-of-distribution data when data is noisy or contains nuisance features.
problem Correlation between in-distribution and out-of-distribution accuracy in noisy or feature-rich data.
method Analyzes the impact of noise and nuisance features on model performance.
result Accuracy on in-distribution and out-of-distribution data can become negatively correlated in noisy or feature-rich data.
Unified study of out-of-distribution generalization across 172 datasets.
problem Measuring and improving robustness of transfer learning models.
method Collect and fine-tune 31k models on 172 dataset pairs, varying architectures and settings.
result In- and out-of-distribution accuracies increase jointly but their relation is dataset-dependent.
Single deep model detects out-of-distribution data with single forward pass.
problem Detecting out-of-distribution data points in neural networks.
method Deterministic uncertainty quantification (DUQ) using gradient penalty for reliable detection.
result Single model outperforms or matches ensemble methods in out-of-distribution detection.
Combining training and post-training methods improves OOD detection accuracy.
problem Deep networks struggle with OOD detection.
method Divided OOD detection methods into training and post-training, then combined them.
result State-of-the-art results in OOD detection achieved.
Proposes a three-stage debiasing framework to improve out-of-distribution accuracy.
problem Inaccurate uncertainty estimations in bias-only models damage ensemble-based debiasing performance.
method Calibrates the bias-only model to improve its uncertainty estimations, creating a three-stage ensemble-based debiasing framework.
result The three-stage debiasing framework consistently outperforms traditional methods in out-of-distribution accuracy.
EntProp increases entropy of clean samples to generate out-of-distribution data for better DNN performance.
problem Improving deep neural networks' accuracy and robustness to out-of-distribution data.
method High entropy propagation using data augmentation and free adversarial training.
result EntProp achieves higher standard accuracy and robustness with lower training cost.
Calibrated ensembles improve both ID and OOD accuracy in distribution shift.
problem Desired balance between in-distribution and out-of-distribution accuracy.
method Ensemble standard and robust models, calibrating on ID data only.
result ID-calibrated ensembles outperform state-of-the-art methods on multiple datasets.
We introduce two challenging datasets that reliably cause machine learning model performance to substantially degrade. The datasets are collected with a simple adversarial filtration technique to create datasets with limited spurious cues. Our datasets' real-world, unmodified examples transfer to various unseen models …
The paper shows strong correlation between in-distribution and out-of-distribution performance in various machine learning models.
problem Understanding reliability of machine learning systems in unseen environments.
method Empirical analysis of various models and distribution shifts on CIFAR-10, ImageNet, and other datasets.
result Out-of-distribution performance is strongly correlated with in-distribution performance across different models and distribution shifts.
A new loss function improves neural networks' out-of-distribution detection without side effects.
problem Neural networks struggle with out-of-distribution detection due to SoftMax loss issues.
method Proposes IsoMax loss replacing SoftMax loss, maintaining high entropy and fast inferences.
result Significantly improves neural networks' out-of-distribution detection performance.
Paper fine-tunes a simulation-driven estimator to reduce out-of-distribution errors.
problem Out-of-distribution errors in simulation-driven parameter estimators.
method Fine-tuning a Two-Stage estimator to improve accuracy for true parameters outside the sampled range.
result The fine-tuning approach reduces out-of-distribution errors and improves accuracy.
A new OOD detector using an overlap index improves accuracy without high computational costs.
problem Effective OOD detection for machine learning models in open-world scenarios.
method Proposes an overlap index-based confidence score function for OOD detection.
result The proposed method achieves competitive accuracy with lower computational costs compared to state-of-the-art detectors.
Sharp analysis of out-of-distribution error in overparameterized models with importance weights.
problem Understanding and quantifying the degradation of performance in overparameterized models when faced with underrepresented data.
method Sharp analysis of an overparameterized Gaussian mixture model with spurious features and cost-sensitive interpolating solutions incorporating importance weights.
result Characterization of a novel tradeoff between worst-case robustness and average accuracy as a function of importance weight magnitude.
Deconfounds neural network representation similarity metrics to improve consistency and accuracy.
problem Confounding by population structure in similarity metrics like RSA and CKA.
method Covariate adjustment regression to adjust for confounders.
result Improves detection of semantically similar neural networks and consistency in transfer learning.
Soft labeling impacts OOD detection in neural networks.
problem Impact of soft labeling on OOD detection in deep neural networks.
method Empirical analysis of how soft labeling affects OOD detection performance.
result Soft labeling can either improve or deteriorate OOD detection performance.
Paper introduces Influence Function to assess OOD generalization stability.
problem Assessing OOD generalization accuracy when target domains are unknown.
method Introduced Influence Function from robust statistics to monitor model stability.
result Accuracy on test domains and Influence Function variance can distinguish OOD algorithms and generalization quality.
Packed-Ensembles improve uncertainty estimation in constrained hardware.
problem Hardware limitations restrict the size of ensembles and network capacity, degrading performance.
method Packed-Ensembles (PE) design and train lightweight structured ensembles by modulating encoding space and parallelizing into a single backbone.
result PE accurately preserves diversity and maintains performance on key metrics like accuracy, calibration, and out-of-distribution detection.
New framework uses geometry of embeddings to predict robustness.
problem Monitoring robustness in models without OOD labels.
method Constructs graphs from embeddings, measures spectral complexity and curvature.
result Representation geometry predicts robustness reliably.
Adaptive Misinformation defends against model stealing attacks by sending incorrect predictions for OOD queries.
problem Model stealing attacks clone target models using black-box query access and a surrogate dataset.
method Selective sending of incorrect predictions for Out-Of-Distribution (OOD) queries to degrade attacker's clone model accuracy.
result Our defense reduces attacker's clone model accuracy by up to 40% while maintaining benign user accuracy under 0.5%.
Study compares data-driven vs model-based MRS quantification strategies, focusing on resilience to out-of-distribution effects.
problem Resilience to out-of-distribution effects in data-driven MRS quantification.
method Compared three data-driven strategies (supervised regression, self-supervised learning, test-time adaptation) against model-based fitting tools.
result Test-time adaptation proved most resilient to out-of-distribution effects, while self-supervised learning achieved intermediate performance.
This paper benchmarks OoDD methods for medical imaging.
problem Medical models trained for one domain may fail on images from a different domain.
method Defined 3 categories of OoD examples and benchmarked methods in 3 medical imaging domains.
result Simple binary classifier on feature representation yields best accuracy and AUPRC.
Study shows latent space OOD detection isn't a reliable proxy for model performance.
problem Evaluating and interpreting deep learning systems on real-world data.
method Empirical investigation of latent space OOD detection and classification accuracy using SAR datasets.
result OOD detection cannot be used as a proxy measure for model performance.
Out-of-distribution (OOD) detection approaches usually present special requirements (e.g., hyperparameter validation, collection of outlier data) and produce side effects (e.g., classification accuracy drop, slower energy-inefficient inferences). We argue that these issues are a consequence of the SoftMax loss anisotro…
Paper presents a method to detect out-of-distribution spectra in intra-operative functional imaging.
problem Detecting out-of-distribution (OoD) spectra in multispectral optical imaging during surgery.
method Information theory-based approach using WAIC with an ensemble of INNs.
result The method is effective in detecting OoD spectra, improving the reliability of functional imaging.
Paper aims to ensure reliable detection of out-of-distribution data with certifiable worst-case guarantees.
problem Deep neural networks are overconfident with OOD inputs, posing safety risks.
method Enforces low confidence and bounds in an l∞-ball around OOD points using interval bound propagation (IBP). result Certifiable worst-case guarantees for OOD detection are possible without significant loss in accuracy.
ATC predicts target domain accuracy using only labeled and unlabeled data.
problem Predicting out-of-distribution performance with limited labeled data.
method Average Thresholded Confidence (ATC) method that learns a threshold on model confidence.
result ATC outperforms previous methods across various types of distribution shifts and datasets.
MixMOOD improves SSDL by selecting unlabelled data based on deep feature similarity.
problem Class distribution mismatch in semi-supervised learning.
method MixMOOD uses deep dataset dissimilarity measures to select unlabelled data.
result MixMOOD selects unlabelled data based on strong correlation with MixMatch accuracy.
Framework extends neural operators to handle functions outside training set.
problem Robust handling of functions beyond the training set.
method Kernel approximation techniques and Reproducing Kernel Hilbert Spaces (RKHSs) theory.
result Theoretical framework and empirical validation for reliable function extension.
Personalized activity recognition improves performance for diverse users.
problem Poor performance of impersonal algorithms for individual users.
method Personalized activity recognition using deep embeddings from a fully convolutional neural network with triplet loss.
result Novel subject triplet loss provides the best performance overall.
Score-based generative models achieve state-of-the-art classification accuracy on CIFAR-10.
problem Improving classification accuracy on natural image datasets.
method Score-based generative models applied to CIFAR-10.
result Score-based models achieve state-of-the-art classification accuracy on CIFAR-10.
PFDL improves deep learning models' OOD generalization by decorrelating feature embeddings.
problem Out-of-distribution generalization in deep learning models.
method PFDL algorithm that optimizes feature decomposition network and image classification model.
result PFDL improves the accuracy of image classification models on OOD datasets.
p-DkNN uses deep representations to detect out-of-distribution data with statistical tests.
problem Lack of reliable confidence estimates in neural networks for safety-critical applications.
method Statistical testing of deep neural network's intermediate hidden representations.
result p-DkNN enables more accurate and reliable predictions by abstaining from incorrect predictions.
New method calibrates deep models for both in-distribution and out-of-distribution samples.
problem Ensuring calibration for deep models in safety-critical applications, especially in OOD regions.
method Geodesic distance and Gaussian kernel to calibrate deep models.
result Proposed KDF and KDN methods achieve well-calibrated posteriors for both in-distribution and out-of-distribution samples.
Study identifies failure modes of machine learning models in out-of-distribution settings.
problem Machine learning models fail to generalize well to new, unseen data.
method Theoretical study of gradient-descent-trained linear classifiers on easy-to-learn tasks, followed by experiments on modern neural networks.
result Two failure modes of spurious correlations are uncovered: geometric and statistical.
Proposes a new method for deep ensembles that improves accuracy and calibration.
problem Improving accuracy and calibration of deep ensembles.
method Estimates confusion matrices of ensemble members and weighs them according to their inferred performance.
result Empirically shows superiority of soft Dawid Skene over ensemble averaging.
Deep neural networks have achieved remarkable success in various challenging tasks. However, the black-box nature of such networks is not acceptable to critical applications, such as healthcare. In particular, the existence of adversarial examples and their overgeneralization to irrelevant, out-of-distribution inputs w…
SVD-RND detects blurred images better than conventional methods.
problem Blurred images can fool conventional OOD detection schemes.
method Constructs a novel RND-based detector that uses blurred images during training.
result SVD-RND outperforms baseline detectors in various domains.
RIA method improves OoD generalization for covariate shift.
problem Improving out-of-distribution generalization under covariate shift.
method Adversarial label invariant graph data augmentations for OoD generalization.
result RIA method achieves high accuracy compared to OoD baselines.
E-Stitchup augments pre-trained embeddings for better model performance and confidence calibration.
problem Improving classification accuracy and confidence calibration of models using pre-trained embeddings.
method Data augmentation methods inspired by Mixup combined with label softening.
result Significantly increased classification accuracy and reduced training time.
Method enhances anomaly detection using contrastive learning and out-of-distribution data.
problem Improving anomaly detection in datasets with limited out-of-distribution data.
method Proposes a contrastive learning method that incorporates out-of-distribution data to enhance anomaly detection performance.
result The method significantly improves anomaly detection performance, even with limited out-of-distribution data.
Bayesian deep ensembles improve prediction accuracy in various settings.
problem Improving prediction accuracy of deep ensembles in out-of-distribution settings.
method Introducing a randomised, untrainable function to each ensemble member, enabling a posterior predictive distribution interpretation.
result Bayesian deep ensembles make more conservative predictions and outperform standard ensembles in various tasks.
Simple methods combine statistical tests for out-of-distribution detection.
problem Detecting data points not following the training distribution.
method Combining classical parametric tests (Rao's score test) and a typicality test.
result Combining Fisher's method of test statistics improves out-of-distribution detection accuracy.
We introduce a measure to quantify ambiguity in deep learning models, improving their reliability.
problem Deep learning models make mistakes on seemingly trivial cases and fail in recognizing what they don't know.
method We define ambiguity based on decision boundaries and convex hulls in feature space, developing a theoretical framework to identify unknowns.
result A single ambiguity measure can detect a significant portion of model mistakes, including adversarial and out-of-distribution inputs.
DRGO improves graph recommendation by mitigating noisy samples and enhancing out-of-distribution generalization.
problem Noisy samples in training data diminish recommendation systems' out-of-distribution generalization.
method DRGO uses a diffusion paradigm to reduce noisy effects and entropy regularization to avoid extreme weights.
result DRGO outperforms current methods in OOD recommendation across various distribution shifts.
Neural network accuracy improves with denser training samples.
problem Improving neural network accuracy on unseen test samples.
method Bounding empirical training error smoothed across activation regions and using it to discard high-risk test samples.
result Discarding high-risk test samples based on error bounds improves prediction accuracy by up to 20%.
A new framework evaluates model performance on single input points, revealing insights into data and model structure.
problem Traditional evaluation methods in machine learning are insufficient for understanding model performance and data structure.
method Developed a pointwise framework to measure model performance on individual input points, analyzing the relationship between pointwise and average performance.
result Profiles of data points reveal different types of correlations between pointwise and average performance, challenging existing models of learning.
Bayesian Gaussian Processes layer detects out-of-distribution data in medical imaging.
problem Detecting out-of-distribution data in medical imaging tasks.
method Parameter-efficient hierarchical convolutional Gaussian Processes in Wasserstein-2 space.
result Uncertainty estimates enable superior out-of-distribution detection compared to previous methods.
Study improves CNN medical image segmentation accuracy and reliability.
problem Over-confident predictions and silent failures in out-of-distribution data.
method Multi-task learning and spectral analysis of CNN feature maps.
result Joint multi-task learning models outperform dedicated models and detect OOD data more accurately.