Scaling for Training Time and Post-hoc Out-of-distribution Detection Enhancement

2023/09/30 by Kai Xu, Rongyu Chen, Xu, Kai +5 · 20 citations
Computer Science · Medicine · #Anomaly Detection Techniques and Applications #Advanced Neural Network Applications #COVID-19 diagnosis using AI

paper · pdf · doi:10.48550/arxiv.2310.00227

Abstract

The capacity of a modern deep learning system to determine if a sample falls within its realm of knowledge is fundamental and important. In this paper, we offer insights and analyses of recent state-of-the-art out-of-distribution (OOD) detection methods - extremely simple activation shaping (ASH). We demonstrate that activation pruning has a detrimental effect on OOD detection, while activation scaling enhances it. Moreover, we propose SCALE, a simple yet effective post-hoc network enhancement method for OOD detection, which attains state-of-the-art OOD detection performance without compromising in-distribution (ID) accuracy. By integrating scaling concepts into the training process to capture a sample's ID characteristics, we propose Intermediate Tensor SHaping (ISH), a lightweight method for training time OOD detection enhancement. We achieve AUROC scores of +1.85% for near-OOD and +0.74% for far-OOD datasets on the OpenOOD v1.5 ImageNet-1K benchmark. Our code and models are available at https://github.com/kai422/SCALE.

Cited by

• Out-of-Distribution Detection for Continual Learning: Design Principles and Benchmarking
• Domain Feature Collapse: Implications for Out-of-Distribution Detection and Solutions
• Improving Out-of-Distribution Detection via Dynamic Covariance Calibration
• Nearest-Class Mean and Logits Agreement for Wildlife Open-Set Recognition
• A Geometry-Based View of Mahalanobis OOD Detection
• A Machine Learning Perspective on Automated Driving Corner Cases
• DynaSubVAE: Adaptive Subgrouping for Scalable and Robust OOD Detection
• ICONIC-444: A 3.1-Million-Image Dataset for OOD Detection Research
• Activation Subspaces for Out-of-Distribution Detection
• Multi-Method Ensemble for Out-of-Distribution Detection
• Retrieval-Augmented Prompt for OOD Detection
• COSTARR: Consolidated Open Set Technique with Attenuation for Robust Recognition
• COOkeD: Ensemble-based OOD detection in the era of zero-shot CLIP
• Out-of-Distribution Detection with Adaptive Top-K Logits Integration
• Spurious-Aware Prototype Refinement for Reliable Out-of-Distribution Detection
• Reliable Few-shot Learning under Dual Noises
• Balanced Hyperbolic Embeddings Are Natural Out-of-Distribution Detectors
• LoRA-BAM: Input Filtering for Fine-tuned LLMs via Boxed Abstraction Monitors over LoRA Layers
• Buffer-free Class-Incremental Learning with Out-of-Distribution Detection
• Kernel PCA for Out-of-Distribution Detection: Non-Linear Kernel Selections and Approximations

Related

• Extremely Simple Activation Shaping for Out-of-Distribution Detection
• AdaSCALE: Adaptive Scaling for OOD Detection
• Logit Scaling for Out-of-Distribution Detection
• Towards Realistic Out-of-Distribution Detection: A Novel Evaluation Framework for Improving Generalization in OOD Detection
• Image Background Serves as Good Proxy for Out-of-distribution Data