Don't Collapse Your Features: Why CenterLoss Hurts OOD Detection and Multi-Scale Mahalanobis Wins

Fuente: arXiv

Saved in:

Bibliographic Details
Main Author:	Ray, Rahul D
Format:	Preprint
Published:	2026
Subjects:	Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Mahalanobis++: Improving OOD Detection via Feature Normalization
by: Mueller, Maximilian, et al.
Published: (2025)

MultiOOD: Scaling Out-of-Distribution Detection for Multiple Modalities
by: Dong, Hao, et al.
Published: (2024)

VOLTA: The Surprising Ineffectiveness of Auxiliary Losses for Calibrated Deep Learning
by: Ray, Rahul D, et al.
Published: (2026)

A Geometry-Based View of Mahalanobis OOD Detection
by: Janiak, Denis, et al.
Published: (2025)

A Geometric Explanation of the Likelihood OOD Detection Paradox
by: Kamkari, Hamidreza, et al.
Published: (2024)

CMOOD: Concept-based Multi-label OOD Detection
by: Liu, Zhendong, et al.
Published: (2024)

Diversify, Don't Fine-Tune: Scaling Up Visual Recognition Training with Synthetic Images
by: Yu, Zhuoran, et al.
Published: (2023)

Proto-OOD: Enhancing OOD Object Detection with Prototype Feature Similarity
by: Chen, Junkun, et al.
Published: (2024)

Feature Purified Transformer With Cross-level Feature Guiding Decoder For Multi-class OOD and Anomaly Deteciton
by: Lin, Jerry Chun-Wei, et al.
Published: (2024)

Don't Play Favorites: Minority Guidance for Diffusion Models
by: Um, Soobin, et al.
Published: (2023)

Don't Judge by the Look: Towards Motion Coherent Video Representation
by: Zhang, Yitian, et al.
Published: (2024)

Scalable Ensemble Diversification for OOD Generalization and Detection
by: Rubinstein, Alexander, et al.
Published: (2024)

Don't Get Me Wrong: How to Apply Deep Visual Interpretations to Time Series
by: Loeffler, Christoffer, et al.
Published: (2022)

Don't Blame the Annotator: Bias Already Starts in the Annotation Instructions
by: Parmar, Mihir, et al.
Published: (2022)

Vision Transformers Don't Need Trained Registers
by: Jiang, Nick, et al.
Published: (2025)

Shadows Don't Lie and Lines Can't Bend! Generative Models don't know Projective Geometry...for now
by: Sarkar, Ayush, et al.
Published: (2023)

Toxicity Assessment in Preclinical Histopathology via Class-Aware Mahalanobis Distance for Known and Novel Anomalies
by: Graf, Olga, et al.
Published: (2026)

AdaNeg: Adaptive Negative Proxy Guided OOD Detection with Vision-Language Models
by: Zhang, Yabin, et al.
Published: (2024)

LAPT: Label-driven Automated Prompt Tuning for OOD Detection with Vision-Language Models
by: Zhang, Yabin, et al.
Published: (2024)

Don't Fear Peculiar Activation Functions: EUAF and Beyond
by: Wang, Qianchao, et al.
Published: (2024)

What Helps---and What Hurts: Bidirectional Explanations for Vision Transformers
by: Su, Qin, et al.
Published: (2026)

Feature-Space Semantic Invariance: Enhanced OOD Detection for Open-Set Domain Generalization
by: Wang, Haoliang, et al.
Published: (2024)

When Robots Should Say "I Don't Know": Benchmarking Abstention in Embodied Question Answering
by: Wu, Tao, et al.
Published: (2025)

Activation Matters: Test-time Activated Negative Labels for OOD Detection with Vision-Language Models
by: Zhang, Yabin, et al.
Published: (2026)

Diversifying Deep Ensembles: A Saliency Map Approach for Enhanced OOD Detection, Calibration, and Accuracy
by: Dereka, Stanislav, et al.
Published: (2023)

Retrieval-Augmented Prompt for OOD Detection
by: Han, Ruisong, et al.
Published: (2025)

When Text and Images Don't Mix: Bias-Correcting Language-Image Similarity Scores for Anomaly Detection
by: Goodge, Adam, et al.
Published: (2024)

Multi-Task Learning with Multi-Annotation Triplet Loss for Improved Object Detection
by: Zhou, Meilun, et al.
Published: (2025)

Prototype Fusion: A Training-Free Multi-Layer Approach to OOD Detection
by: Gul, Shreen, et al.
Published: (2026)

Continual Learning: Forget-free Winning Subnetworks for Video Representations
by: Kang, Haeyong, et al.
Published: (2023)

Don't Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs
by: Jiao, Pengkun, et al.
Published: (2025)

Render, Don't Decode: Weight-Space World Models with Latent Structural Disentanglement
by: Nzoyem, Roussel Desmond, et al.
Published: (2026)

Don't Blind Your VLA: Aligning Visual Representations for OOD Generalization
by: Kachaev, Nikita, et al.
Published: (2025)

Distribution Shifts at Scale: Out-of-distribution Detection in Earth Observation
by: Ekim, Burak, et al.
Published: (2024)

COOkeD: Ensemble-based OOD detection in the era of zero-shot CLIP
by: Humblot-Renaux, Galadrielle, et al.
Published: (2025)

Focus, Don't Prune: Identifying Instruction-Relevant Regions for Information-Rich Image Understanding
by: Kwon, Mincheol, et al.
Published: (2026)

Don't Miss the Forest for the Trees: Attentional Vision Calibration for Large Vision Language Models
by: Woo, Sangmin, et al.
Published: (2024)

Large Pre-Training Datasets Don't Always Guarantee Robustness after Fine-Tuning
by: Hwang, Jaedong, et al.
Published: (2024)

Don't trust your eyes: on the (un)reliability of feature visualizations
by: Geirhos, Robert, et al.
Published: (2023)

I Detect What I Don't Know: Incremental Anomaly Learning with Stochastic Weight Averaging-Gaussian for Oracle-Free Medical Imaging
by: Yadav, Nand Kumar, et al.
Published: (2025)