TIE: A Training-Inversion-Exclusion Framework for Visually Interpretable and Uncertainty-Guided Out-of-Distribution Detection

TL;DR

Motivates the need for unified uncertainty estimation and OOD detection without external OOD data. Proposes TIE, which integrates network inversion into the training loop and uses a garbage class with uncertainty-guided, iterative exclusion to refine decision boundaries. Demonstrates two-tier inference—direct OOD rejection via the garbage class and threshold-based fine-grained OOD detection—achieving near-zero FPR@95%TPR on MNIST and FashionMNIST and robust OOD discrimination across diverse datasets. The work provides interpretable visualizations of class manifolds and calibrated predictive uncertainty, offering a self-contained framework for reliable anomaly detection in safety-critical settings.

Abstract

Deep neural networks often struggle to recognize when an input lies outside their training experience, leading to unreliable and overconfident predictions. Building dependable machine learning systems therefore requires methods that can both estimate predictive \textit{uncertainty} and detect \textit{out-of-distribution (OOD)} samples in a unified manner. In this paper, we propose \textbf{TIE: a Training--Inversion--Exclusion} framework for visually interpretable and uncertainty-guided anomaly detection that jointly addresses these challenges through iterative refinement. TIE extends a standard $n$-class classifier to an $(n+1)$-class model by introducing a garbage class initialized with Gaussian noise to represent outlier inputs. Within each epoch, TIE performs a closed-loop process of \textit{training, inversion, and exclusion}, where highly uncertain inverted samples reconstructed from the just-trained classifier are excluded into the garbage class. Over successive iterations, the inverted samples transition from noisy artifacts into visually coherent class prototypes, providing transparent insight into how the model organizes its learned manifolds. During inference, TIE rejects OOD inputs by either directly mapping them to the garbage class or producing low-confidence, uncertain misclassifications within the in-distribution classes that are easily separable, all without relying on external OOD datasets. A comprehensive threshold-based evaluation using multiple OOD metrics and performance measures such as \textit{AUROC}, \textit{AUPR}, and \textit{FPR@95\%TPR} demonstrates that TIE offers a unified and interpretable framework for robust anomaly detection and calibrated uncertainty estimation (UE) achieving near-perfect OOD detection with \textbf{\(\!\approx\!\) 0 FPR@95\%TPR} when trained on MNIST or FashionMNIST and tested against diverse unseen datasets.

Figures (6)

• Figure 1: Overview of the proposed TIE framework.
• Figure 2: Evolution of inverted samples across epochs for select classes in the TIE framework.
• Figure 3: Evolution of averaged metrics associated with the inverted samples across epochs for all classes in MNIST.
• Figure 4: t-SNE visualization of features in the latent space across epochs for all $(n+1)$ classes of MNIST shows compact in-distribution clusters surrounding the expanding garbage class region.
• Figure 5: Threshold-based OOD Detection on MNIST. Top: AUROC and AUPR curves for Mahalanobis distance across multiple OOD datasets. Bottom: Score separation plots showing distribution of Mahalanobis distances for in-distribution (blue) and OOD (red) samples.