Monitizer: Automating Design and Evaluation of Neural Network
Monitors

D 2024

Monitizer: Automating Design and Evaluation of Neural Network Monitors

AZEEM, Muqsit; Marta GROBELNA; Sudeep KANAV; Jan KŘETÍNSKÝ; Stefanie MOHR et. al.

Základní údaje

Originální název

Monitizer: Automating Design and Evaluation of Neural Network Monitors

Autoři

AZEEM, Muqsit; Marta GROBELNA; Sudeep KANAV; Jan KŘETÍNSKÝ; Stefanie MOHR a Sabine RIEDER

Vydání

Montreal, Canada, CAV 2024, International Conference on Computer Aided Verification, od s. 265-279, 15 s. 2024

Nakladatel

Springer

Další údaje

Jazyk

angličtina

Typ výsledku

Stať ve sborníku

Obor

10201 Computer sciences, information science, bioinformatics

Stát vydavatele

Německo

Utajení

není předmětem státního či obchodního tajemství

Forma vydání

tištěná verze "print"

Impakt faktor

Impact factor: 0.402 v roce 2005

Kód RIV

RIV/00216224:14330/24:00136698

Organizační jednotka

Fakulta informatiky

ISBN

978-3-031-65629-3

ISSN

DOI

https://doi.org/10.1007/978-3-031-65630-9_14

UT WoS

001307890400014

EID Scopus

2-s2.0-85200663897

Klíčová slova anglicky

Neural Networks; Monitoring; Hyperparameter Tuning

Štítky

core_A, firank_1

Příznaky

Mezinárodní význam, Recenzováno

Změněno: 4. 4. 2025 12:03, RNDr. Pavel Šmerk, Ph.D.

Anotace

V originále

The behavior of neural networks (NNs) on previously unseen types of data (out-of-distribution or OOD) is typically unpredictable. This can be dangerous if the network's output is used for decision making in a safety-critical system. Hence, detecting that an input is OOD is crucial for the safe application of the NN. Verification approaches do not scale to practical NNs, making runtime monitoring more appealing for practical use. While various monitors have been suggested recently, their optimization for a given problem, as well as comparison with each other and reproduction of results, remain challenging.

Návaznosti

MUNI/A/1081/2022, interní kód MU

Název: Modelování, analýza a verifikace (2023)

Investor: Masarykova univerzita, Modelování, analýza a verifikace (2023)

MUNI/A/1433/2022, interní kód MU

Název: Zapojení studentů Fakulty informatiky do mezinárodní vědecké komunity 23

Investor: Masarykova univerzita, Zapojení studentů Fakulty informatiky do mezinárodní vědecké komunity 23

MUNI/I/1757/2021, interní kód MU

Název: MUNI Award in Science and Humanities (Akronym: Křetínský)

Investor: Masarykova univerzita, MUNI Award in Science and Humanities, MASH - MUNI Award in Science and Humanities

Citovat

AZEEM, Muqsit; Marta GROBELNA; Sudeep KANAV; Jan KŘETÍNSKÝ; Stefanie MOHR a Sabine RIEDER. Monitizer: Automating Design and Evaluation of Neural Network Monitors. In CAV 2024, International Conference on Computer Aided Verification. Montreal, Canada: Springer, 2024, s. 265-279. ISBN 978-3-031-65629-3. Dostupné z: https://doi.org/10.1007/978-3-031-65630-9_14.

@inproceedings{2421707,
   author = {Azeem, Muqsit and Grobelna, Marta and Kanav, Sudeep and Křetínský, Jan and Mohr, Stefanie and Rieder, Sabine},
   address = {Montreal, Canada},
   booktitle = {CAV 2024, International Conference on Computer Aided Verification},
   doi = {https://doi.org/10.1007/978-3-031-65630-9_14},
   keywords = {Neural Networks; Monitoring; Hyperparameter Tuning},
   howpublished = {tištěná verze "print"},
   language = {eng},
   location = {Montreal, Canada},
   isbn = {978-3-031-65629-3},
   pages = {265-279},
   publisher = {Springer},
   title = {Monitizer: Automating Design and Evaluation of Neural Network Monitors},
   year = {2024}
}

TY  - CONF
ID  - 2421707
AU  - Azeem, Muqsit - Grobelna, Marta - Kanav, Sudeep - Křetínský, Jan - Mohr, Stefanie - Rieder, Sabine
PY  - 2024
TI  - Monitizer: Automating Design and Evaluation of Neural Network Monitors
PB  - Springer
CY  - Montreal, Canada
SN  - 9783031656293
KW  - Neural Networks
KW  - Monitoring
KW  - Hyperparameter Tuning
N2  - The behavior of neural networks (NNs) on previously unseen types of data (out-of-distribution or OOD) is typically unpredictable. This can be dangerous if the network's output is used for decision making in a safety-critical system. Hence, detecting that an input is OOD is crucial for the safe application of the NN. Verification approaches do not scale to practical NNs, making runtime monitoring more appealing for practical use. While various monitors have been suggested recently, their optimization for a given problem, as well as comparison with each other and reproduction of results, remain challenging.
ER  -

AZEEM, Muqsit; Marta GROBELNA; Sudeep KANAV; Jan KŘETÍNSKÝ; Stefanie MOHR a Sabine RIEDER. Monitizer: Automating Design and Evaluation of Neural Network Monitors. In \textit{CAV 2024, International Conference on Computer Aided Verification}. Montreal, Canada: Springer, 2024, s.~265-279. ISBN~978-3-031-65629-3. Dostupné z: https://doi.org/10.1007/978-3-031-65630-9\_{}14.

Přehled o publikaci