Proofs depending on the axiom of choice show a geometric signature in neural embeddings of tactic sequences that weakens with dependency-graph distance and correlates with prover failure rates.
hub
Enhancing the reliability of out-of-distribution image detection in neural networks
27 Pith papers cite this work, alongside 651 external citations. Polarity classification is still indexing.
27
Pith papers citing it
651
external citations · Pith
abstract
We consider the problem of detecting out-of-distribution images in neural networks. We propose ODIN, a simple and effective method that does not require any change to a pre-trained neural network. Our method is based on the observation that using temperature scaling and adding small perturbations to the input can separate the softmax score distributions between in- and out-of-distribution images, allowing for more effective detection. We show in a series of experiments that ODIN is compatible with diverse network architectures and datasets. It consistently outperforms the baseline approach by a large margin, establishing a new state-of-the-art performance on this task. For example, ODIN reduces the false positive rate from the baseline 34.7% to 4.3% on the DenseNet (applied to CIFAR-10) when the true positive rate is 95%.
hub tools
citation-role summary
background 3
citation-polarity summary
roles
background 3polarities
background 3representative citing papers
CURE-OOD is the first benchmark for evaluating OOD detection in survival prediction under controlled CT acquisition shifts, showing that standard detectors often fail and providing a survival-aware baseline.
Sparse autoencoders on ViT class tokens reveal stable Class Activation Profiles for in-distribution data, enabling OOD detection via divergence from core energy profiles.
SLE-FNO achieves zero forgetting and strong plasticity-stability balance in continual learning for FNO surrogate models of pulsatile blood flow by adding minimal single-layer extensions across four out-of-distribution tasks.
HyperNSD models hypergraph node states as an incidence-aware SDE whose pathwise variability yields competitive uncertainty estimates for OOD and misclassification detection.
ASTRANet combines a redshift-free spectral classifier, a 16-score anomaly detector, and conformal prediction to identify and calibrate uncertainty for out-of-taxonomy astronomical transients.
Proposes using sparse autoencoders to extract class-conditioned concept vectors, then measuring logit stability under targeted perturbations as an interpretable OOD signal for deep networks in medical imaging.
Debiased negative mining via Monte-Carlo sampling from ID labels and unlabeled wild data improves OOD detection with VLMs and achieves new state-of-the-art results.
RS-GNNs predict random sets over classes using belief functions to jointly produce class probabilities and epistemic uncertainty estimates for graph nodes.
Multi-layer SAE transitions capture domain-specific signatures that distinguish OOD texts in Gemma-2 models.
HCM estimates uncertainty in neural network outputs by quantifying violation of a unit hypersphere constraint on the normalized direction vector.
Avoiding CenterLoss improves OOD detection via multi-scale Mahalanobis on L2-normalized features, yielding 0.9483 AUROC on CIFAR-10 while preserving competitive in-distribution accuracy.
ConjNorm reframes OOD detection score design as optimizing norm p in an exponential family density model via a Bregman divergence theorem, with a tractable Monte Carlo estimator, claiming SOTA gains on CIFAR-100 and ImageNet-1K.
PointCaM proposes a cut-and-mix mechanism with an Unknown-Point Simulator and Estimator to improve open-set recognition on point clouds by simulating out-of-distribution data and using multi-level features.
Language models show good calibration when asked to estimate the probability that their own answers are correct, with performance improving as models get larger.
DFDNet disentangles content from style via dual modules to boost fine-grained OOD detection performance on multiple datasets.
Hyperspherical time-frequency representations learned via von Mises-Fisher likelihood improve OOD detection on UCR and UEA archives using k-NN and Mahalanobis scores over contrastive baselines.
Proposes TCD, a three-step conditional diffusion model with ICD module, claiming superior fidelity and generalization for LFM 3D reconstruction.
RefCal uses a supervised contrastive loss to promote refinement alongside calibration in DNN training, reporting better accuracy, refinement, and ECE than baselines on imbalanced CIFAR-100-LT.
UniAlign improves robustness of deep learning NTC models under distribution shifts via domain alignment fine-tuning and stable ensembling, yielding 2.51% accuracy and 2.71% F1 gains over standard training on three public datasets.
HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.
ODiSAR uses a Transformer digital twin with reconstruction error and Monte Carlo dropout to detect OOD events in self-adaptive robots, reporting up to 98% AUROC on office navigation and maritime ship tasks.
Vision foundation model embeddings with density modeling outperform state-of-the-art methods for unsupervised semantic and covariate shift detection in autonomous driving inputs.
Developed 19 metamorphic relations to test correlation detection and LSTM forecasting in an outage prediction application, uncovering 8 unknown issues in the live system and detecting 65.9% of injected bugs via mutation testing.
citing papers explorer
-
Geometric Measurements of the Axiom of Choice in Neural Proof Embeddings
Proofs depending on the axiom of choice show a geometric signature in neural embeddings of tactic sequences that weakens with dependency-graph distance and correlates with prover failure rates.
-
CURE-OOD: Benchmarking Out-of-Distribution Detection for Survival Prediction
CURE-OOD is the first benchmark for evaluating OOD detection in survival prediction under controlled CT acquisition shifts, showing that standard detectors often fail and providing a survival-aware baseline.
-
Sparsity as a Key: Unlocking New Insights from Latent Structures for Out-of-Distribution Detection
Sparse autoencoders on ViT class tokens reveal stable Class Activation Profiles for in-distribution data, enabling OOD detection via divergence from core energy profiles.
-
SLE-FNO: Single-Layer Extensions for Task-Agnostic Continual Learning in Fourier Neural Operators
SLE-FNO achieves zero forgetting and strong plasticity-stability balance in continual learning for FNO surrogate models of pulsatile blood flow by adding minimal single-layer extensions across four out-of-distribution tasks.
-
Hypergraph Neural Stochastic Diffusion: An SDE Framework for Uncertainty Estimation
HyperNSD models hypergraph node states as an incidence-aware SDE whose pathwise variability yields competitive uncertainty estimates for OOD and misclassification detection.
-
Catching Disguised Transients with ASTRANet: Anomaly-Aware Spectroscopic Classification and Conformal Calibration
ASTRANet combines a redshift-free spectral classifier, a 16-score anomaly detector, and conformal prediction to identify and calibrate uncertainty for out-of-taxonomy astronomical transients.
-
When Confidence Lacks Concepts: Interpretable OOD Detection via Representation Perturbations
Proposes using sparse autoencoders to extract class-conditioned concept vectors, then measuring logit stability under targeted perturbations as an interpretable OOD signal for deep networks in medical imaging.
-
Debiased Negative Mining Improves Out-of-distribution Detection with Pre-trained Vision-Language Models
Debiased negative mining via Monte-Carlo sampling from ID labels and unlabeled wild data improves OOD detection with VLMs and achieves new state-of-the-art results.
-
Random-Set Graph Neural Networks
RS-GNNs predict random sets over classes using belief functions to jointly produce class probabilities and epistemic uncertainty estimates for graph nodes.
-
Domain Restriction via Multi SAE Layer Transitions
Multi-layer SAE transitions capture domain-specific signatures that distinguish OOD texts in Gemma-2 models.
-
Uncertainty Estimation via Hyperspherical Confidence Mapping
HCM estimates uncertainty in neural network outputs by quantifying violation of a unit hypersphere constraint on the normalized direction vector.
-
Don't Collapse Your Features: Why CenterLoss Hurts OOD Detection and Multi-Scale Mahalanobis Wins
Avoiding CenterLoss improves OOD detection via multi-scale Mahalanobis on L2-normalized features, yielding 0.9483 AUROC on CIFAR-10 while preserving competitive in-distribution accuracy.
-
ConjNorm: Tractable Density Estimation for Out-of-Distribution Detection
ConjNorm reframes OOD detection score design as optimizing norm p in an exponential family density model via a Bregman divergence theorem, with a tractable Monte Carlo estimator, claiming SOTA gains on CIFAR-100 and ImageNet-1K.
-
PointCaM: Cut-and-Mix for Open-Set Point Cloud Learning
PointCaM proposes a cut-and-mix mechanism with an Unknown-Point Simulator and Estimator to improve open-set recognition on point clouds by simulating out-of-distribution data and using multi-level features.
-
Language Models (Mostly) Know What They Know
Language models show good calibration when asked to estimate the probability that their own answers are correct, with performance improving as models get larger.
-
Dual Feature Decoupling for Fine-Grained OOD Detection
DFDNet disentangles content from style via dual modules to boost fine-grained OOD detection performance on multiple datasets.
-
Learning Hyperspherical Time-Frequency Representations for Time-Series Out-of-Distribution Detection
Hyperspherical time-frequency representations learned via von Mises-Fisher likelihood improve OOD detection on UCR and UEA archives using k-NN and Mahalanobis scores over contrastive baselines.
-
Three-Step Conditional Diffusion 3D Reconstruction for Light-Field Microscopy
Proposes TCD, a three-step conditional diffusion model with ICD module, claiming superior fidelity and generalization for LFM 3D reconstruction.
-
Enhancing Deep Neural Network Reliability with Refinement and Calibration
RefCal uses a supervised contrastive loss to promote refinement alongside calibration in DNN training, reporting better accuracy, refinement, and ECE than baselines on imbalanced CIFAR-100-LT.
-
UniAlign: A Model-Agnostic Framework for Robust Network Traffic Classification under Distribution Shifts
UniAlign improves robustness of deep learning NTC models under distribution shifts via domain alignment fine-tuning and stable ensembling, yielding 2.51% accuracy and 2.71% F1 gains over standard training on three public datasets.
-
HEDP: A Hybrid Energy-Distance Prompt-based Framework for Domain Incremental Learning
HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.
-
Out of Distribution Detection in Self-adaptive Robots with AI-powered Digital Twins
ODiSAR uses a Transformer digital twin with reconstruction error and Monte Carlo dropout to detect OOD events in self-adaptive robots, reporting up to 98% AUROC on office navigation and maritime ship tasks.
-
Benchmarking Vision Foundation Models for Input Monitoring in Autonomous Driving
Vision foundation model embeddings with density modeling outperform state-of-the-art methods for unsupervised semantic and covariate shift detection in autonomous driving inputs.
-
Metamorphic Testing of a Deep Learning based Forecaster
Developed 19 metamorphic relations to test correlation detection and LSTM forecasting in an outage prediction application, uncovering 8 unknown issues in the live system and detecting 65.9% of injected bugs via mutation testing.
-
AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensors
AEGIS combines SemantiGAN filtering with evidential learning on five handcrafted instability metrics to detect adversarial attacks, reporting 92.1% AUROC on Tiny ImageNet across six attack types.
-
Out-of-Distribution (OOD) Detectors for Open-Set RF Fingerprinting
Applies OOD detectors to open-set RF fingerprinting via an information-theoretic framework and demonstrates effective tuning without OOD data on the POWDER dataset.
-
Beyond Toy Benchmarks: A Systematic Evaluation of OOD Detection Methods For Plant Pathology Classification
Energy-based fine-tuning outperforms other OOD detection methods on the real-world Plant Pathology 2021 dataset, improving detection over softmax while maintaining in-distribution accuracy.