Gaussian-Based Runtime Detection of Out-of-distribution Inputs for Neural Networks

被引：5

作者：

Hashemi, Vahid ^{[2
]}

Kretinsky, Jan ^{[1
]}

Mohr, Stefanie ^{[1
]}

Seferis, Emmanouil ^{[1
,2
]}

机构：

[1] Tech Univ Munich, Munich, Germany

[2] AUDI AG, Ettingerstr 60, D-85057 Ingolstadt, Germany

来源：

RUNTIME VERIFICATION (RV 2021) | 2021年 / 12974卷

关键词：

D O I：

10.1007/978-3-030-88494-9_14

中图分类号：

TP31 [计算机软件];

学科分类号：

081202 ; 0835 ;

摘要：

In this short paper, we introduce a simple approach for run-time monitoring of deep neural networks and show how to use it for out-of-distribution detection. The approach is based on inferring Gaussian models of some of the neurons and layers. Despite its simplicity, it performs better than recently introduced approaches based on interval abstractions which are traditionally used in verification.

引用

页码：254 / 264

页数：11

共 50 条

[21] On the Learnability of Out-of-distribution Detection
Fang, Zhen
Li, Yixuan
Liu, Feng
Han, Bo
Lu, Jie
Journal of Machine Learning Research, 2024, 25
[22] A Data-centric Framework to Endow Graph Neural Networks with Out-Of-Distribution Detection Ability
Guo, Yuxin
Yang, Cheng
Chen, Yuluo
Liu, Jixi
Shi, Chuan
Du, Junping
PROCEEDINGS OF THE 29TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, KDD 2023, 2023, : 638 - 648
[23] Entropic Out-of-Distribution Detection
Macedo, David
Ren, Tsang Ing
Zanchettin, Cleber
Oliveira, Adriano L., I
Ludermir, Teresa
2021 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), 2021,
[24] Watermarking for Out-of-distribution Detection
Wang, Qizhou
Liu, Feng
Zhang, Yonggang
Zhang, Jing
Gong, Chen
Liu, Tongliang
Han, Bo
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
[25] Is Out-of-Distribution Detection Learnable?
Fang, Zhen
Li, Yixuan
Lu, Jie
Dong, Jiahua
Han, Bo
Liu, Feng
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
[26] On the Learnability of Out-of-distribution Detection
Fang, Zhen
Li, Yixuan
Liu, Feng
Han, Bo
Lu, Jie
JOURNAL OF MACHINE LEARNING RESEARCH, 2024, 25
[27] OODCN: Out-Of-Distribution Detection in Capsule Networks for Fault Identification
Mitiche, Imene
Salimy, Alireza
Werner, Falk
Boreham, Philip
Nesbitt, Alan
Morison, Gordon
29TH EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO 2021), 2021, : 1686 - 1690
[28] Fault Diagnosis of PV Modules based on Convolution Neural Network and Out-of-distribution Detection
Liu, Mengcheng
Hong, Liu
Sheng, Jie
Li, Feng
Zhu, Jin
Ling, Qiang
2023 35TH CHINESE CONTROL AND DECISION CONFERENCE, CCDC, 2023, : 1170 - 1175
[29] WOOD: Wasserstein-Based Out-of-Distribution Detection
Wang, Yinan
Sun, Wenbo
Jin, Jionghua
Kong, Zhenyu
Yue, Xiaowei
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2024, 46 (02) : 944 - 956
[30] High-Efficient Memristor-Based Bayesian Convolutional Neural Networks for Out-of-Distribution Detection by Uncertainty Estimation
Lin, Yudeng
Zhang, Qingtian
Gao, Bin
Tang, Jianshi
Zhao, Han
Qin, Qi
Wang, Ze
Qian, He
Wu, Huaqiang
IEEE TRANSACTIONS ON ELECTRON DEVICES, 2025, 72 (01) : 206 - 214

← 1 2 3 4 5 →