5055710
doi
10.5281/zenodo.5055710
oai:zenodo.org:5055710
user-eu
Paolo Tonella
Università della Svizzera italiana
Fail-Safe Execution of Deep Learning based Systems through Uncertainty Monitoring
Michael Weiss
Università della Svizzera italiana
doi:10.1109/ICST49551.2021.00015
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
<p>Modern software systems rely on Deep Neural Networks (DNN) when processing complex, unstructured inputs, such as images, videos, natural language texts or audio signals. Provided the intractably large size of such input spaces, the intrinsic limitations of learning algorithms and the ambiguity about the expected predictions for some of the inputs, not only there is no guarantee that DNN's predictions are always correct, but rather developers must safely assume a low, though not negligible, error probability. A fail-safe Deep Learning based System (DLS) is one equipped to handle DNN faults by means of a supervisor, capable of recognizing predictions that should not be trusted and that should activate a healing procedure bringing the DLS to a safe state.</p>
<p>In this paper, we propose an approach to use DNN uncertainty estimators to implement such supervisor. We first discuss advantages and disadvantages of existing approaches to measure uncertainty for DNNs and propose novel metrics for the empirical assessment of the supervisor that rely on such approaches. We then describe our publicly available tool Uncertainty-Wizard, which allows transparent estimation of uncertainty for regular tf.keras DNNs. Lastly, we discuss a large-scale study conducted on four different subjects to empirically validate the approach, reporting the lessons-learned as guidance for software engineers who intend to monitor uncertainty for fail-safe execution of DLS.</p>
Zenodo
2020-09-01
info:eu-repo/semantics/report
5055709
user-eu
award_title=Self-assessment Oracles for Anticipatory Testing; award_number=787703; award_identifiers_scheme=url; award_identifiers_identifier=https://cordis.europa.eu/projects/787703; funder_id=00k4n6c32; funder_name=European Commission;
1625190497.897649
903143
md5:ed486374270224a43e2ebc6cf3cd2a5e
https://zenodo.org/records/5055710/files/TR-Precrime-2020-05.pdf
public
10.1109/ICST49551.2021.00015
Is obsoleted by
doi
10.5281/zenodo.5055709
isVersionOf
doi