Into the Unknown : Active Monitoring of Neural Networks

Cite This

Files in this item

Files Size Format View

There are no files associated with this item.

LUKINA, Anna, Christian SCHILLING, Thomas A. HENZINGER, 2021. Into the Unknown : Active Monitoring of Neural Networks. 21st International Conference, RV 2021. Virtual Event, Oct 11, 2021 - Oct 14, 2021. In: FENG, Lu, ed., Dana FISMAN, ed.. Runtime Verification : 21st International Conference, RV 2021, Virtual Event, October 11-14, 2021, proceedings. Cham:Springer, pp. 42-61. ISSN 0302-9743. eISSN 1611-3349. ISBN 978-3-030-88493-2. Available under: doi: 10.1007/978-3-030-88494-9_3

@inproceedings{Lukina2021Unkno-55818, title={Into the Unknown : Active Monitoring of Neural Networks}, year={2021}, doi={10.1007/978-3-030-88494-9_3}, number={12974}, isbn={978-3-030-88493-2}, issn={0302-9743}, address={Cham}, publisher={Springer}, series={Lecture Notes in Computer Science}, booktitle={Runtime Verification : 21st International Conference, RV 2021, Virtual Event, October 11-14, 2021, proceedings}, pages={42--61}, editor={Feng, Lu and Fisman, Dana}, author={Lukina, Anna and Schilling, Christian and Henzinger, Thomas A.} }

<rdf:RDF xmlns:dcterms="" xmlns:dc="" xmlns:rdf="" xmlns:bibo="" xmlns:dspace="" xmlns:foaf="" xmlns:void="" xmlns:xsd="" > <rdf:Description rdf:about=""> <bibo:uri rdf:resource=""/> <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/> <dspace:isPartOfCollection rdf:resource=""/> <foaf:homepage rdf:resource="http://localhost:8080/jspui"/> <dc:creator>Lukina, Anna</dc:creator> <dc:contributor>Lukina, Anna</dc:contributor> <dcterms:title>Into the Unknown : Active Monitoring of Neural Networks</dcterms:title> <dcterms:abstract xml:lang="eng">Neural-network classifiers achieve high accuracy when predicting the class of an input that they were trained to identify. Maintaining this accuracy in dynamic environments, where inputs frequently fall outside the fixed set of initially known classes, remains a challenge. The typical approach is to detect inputs from novel classes and retrain the classifier on an augmented dataset. However, not only the classifier but also the detection mechanism needs to adapt in order to distinguish between newly learned and yet unknown input classes. To address this challenge, we introduce an algorithmic framework for active monitoring of a neural network. A monitor wrapped in our framework operates in parallel with the neural network and interacts with a human user via a series of interpretable labeling queries for incremental adaptation. In addition, we propose an adaptive quantitative monitor to improve precision. An experimental evaluation on a diverse set of benchmarks with varying numbers of classes confirms the benefits of our active monitoring framework in dynamic scenarios.</dcterms:abstract> <dc:creator>Schilling, Christian</dc:creator> <dcterms:available rdf:datatype="">2021-12-09T10:33:44Z</dcterms:available> <dc:contributor>Henzinger, Thomas A.</dc:contributor> <dcterms:isPartOf rdf:resource=""/> <dcterms:issued>2021</dcterms:issued> <dc:contributor>Schilling, Christian</dc:contributor> <dc:creator>Henzinger, Thomas A.</dc:creator> <dc:date rdf:datatype="">2021-12-09T10:33:44Z</dc:date> <dc:language>eng</dc:language> </rdf:Description> </rdf:RDF>

This item appears in the following Collection(s)

Search KOPS


My Account