Deep Neural Networks (DNNs) are nowadays widely used in low-cost accelerators, characterized by limited computational resources. These models, and in particular DNNs for image classification, are becoming increasingly popular in safety-critical applications, where they are required to be highly reliable. Unfortunately, increasing DNNs reliability without computational overheads, which might not be affordable in low-power devices, is a non-trivial task. Our intuition is to detect network executions affected by faults as outliers with respect to the distribution of normal network's output. To this purpose, we propose to exploit Open-Set Recognition (OSR) techniques to perform Fault Detection in an extremely low-cost manner. In particuar, we analyze the Maximum Logit Score (MLS), which is an established Open-Set Recognition technique, and compare it against other well-known OSR methods, namely OpenMax, energy-based out-of-distribution detection and ODIN. Our experiments, performed on a ResNet-20 classifier trained on CIFAR-10 and SVHN datasets, demonstrate that MLS guarantees satisfactory detection performance while adding a negligible computational overhead. Most remarkably, MLS is extremely convenient to configure and deploy, as it does not require any modification or re-training of the existing network. A discussion of the advantages and limitations of the analysed solutions concludes the paper.

Open-Set Recognition: an Inexpensive Strategy to Increase DNN Reliability / Gavarini, Gabriele; Stucchi, Diego; Ruospo, Annachiara; Boracchi, Giacomo; Sanchez, Ernesto. - ELETTRONICO. - (2022). ((Intervento presentato al convegno The 28th IEEE International Symposium on On-Line Testing and Robust System Design (IOLTS 2022) tenutosi a Torino nel September, 12th-14th 2022 [10.1109/IOLTS56730.2022.9897805].

Open-Set Recognition: an Inexpensive Strategy to Increase DNN Reliability

Gavarini, Gabriele;Ruospo, Annachiara;Sanchez, Ernesto
2022

Abstract

Deep Neural Networks (DNNs) are nowadays widely used in low-cost accelerators, characterized by limited computational resources. These models, and in particular DNNs for image classification, are becoming increasingly popular in safety-critical applications, where they are required to be highly reliable. Unfortunately, increasing DNNs reliability without computational overheads, which might not be affordable in low-power devices, is a non-trivial task. Our intuition is to detect network executions affected by faults as outliers with respect to the distribution of normal network's output. To this purpose, we propose to exploit Open-Set Recognition (OSR) techniques to perform Fault Detection in an extremely low-cost manner. In particuar, we analyze the Maximum Logit Score (MLS), which is an established Open-Set Recognition technique, and compare it against other well-known OSR methods, namely OpenMax, energy-based out-of-distribution detection and ODIN. Our experiments, performed on a ResNet-20 classifier trained on CIFAR-10 and SVHN datasets, demonstrate that MLS guarantees satisfactory detection performance while adding a negligible computational overhead. Most remarkably, MLS is extremely convenient to configure and deploy, as it does not require any modification or re-training of the existing network. A discussion of the advantages and limitations of the analysed solutions concludes the paper.
978-1-6654-7355-2
File in questo prodotto:
File Dimensione Formato  
postPrintIOLTS.pdf

non disponibili

Tipologia: 2a Post-print versione editoriale / Version of Record
Licenza: Non Pubblico - Accesso privato/ristretto
Dimensione 323.21 kB
Formato Adobe PDF
323.21 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
Open-Set_Recognition_an_Inexpensive_Strategy_to_Increase_DNN_Reliability.pdf

non disponibili

Tipologia: 2a Post-print versione editoriale / Version of Record
Licenza: Non Pubblico - Accesso privato/ristretto
Dimensione 383.29 kB
Formato Adobe PDF
383.29 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

Caricamento pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/2971537