Towards A Rigorous Evaluation Of XAI Methods On Time Series

Schlegel, Udo; Arnout, Hiba; El-Assady, Mennatallah; Oelke, Daniela; Keim, Daniel A.

Towards A Rigorous Evaluation Of XAI Methods On Time Series

Dateien

Schlegel_2-1iaybk1cefn2y7.pdfGröße: 487.17 KBDownloads: 655

Datum

2019

Autor:innen

Schlegel, Udo

Arnout, Hiba

El-Assady, Mennatallah

Oelke, Daniela

Keim, Daniel A.

Open Access-Veröffentlichung

Open Access Green

Sammlungen

Informatik und Informationswissenschaft

Publikationstyp

Beitrag zu einem Konferenzband

Publikationsstatus

Published

Erschienen in

2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW). Piscataway, NJ: IEEE, 2019, pp. 4321-4325. ISSN 2473-9936. eISSN 2473-9944. ISBN 978-1-72815-023-9. Available under: doi: 10.1109/ICCVW.2019.00516

Zusammenfassung

Explainable Artificial Intelligence (XAI) methods are typically deployed to explain and debug black-box machine learning models. However, most proposed XAI methods are black-boxes themselves and designed for images. Thus, they rely on visual interpretability to evaluate and prove explanations. In this work, we apply XAI methods previously used in the image and text-domain on time series. We present a methodology to test and evaluate various XAI methods on time series by introducing new verification techniques to incorporate the temporal dimension. We further conduct preliminary experiments to assess the quality of selected XAI method explanations with various verification methods on a range of datasets and inspecting quality metrics on it. We demonstrate that in our initial experiments, SHAP works robust for all models, but others like DeepLIFT, LRP, and Saliency Maps work better with specific architectures.

Fachgebiet (DDC)

004 Informatik

Schlagwörter

Time-Series, explainable-ai, explainable-ai-evaluation

Konferenz

2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), 27. Okt. 2019 - 28. Okt. 2019, Seoul, Korea (South)

Zitieren

ISO 690

SCHLEGEL, Udo, Hiba ARNOUT, Mennatallah EL-ASSADY, Daniela OELKE, Daniel A. KEIM, 2019. Towards A Rigorous Evaluation Of XAI Methods On Time Series. 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW). Seoul, Korea (South), 27. Okt. 2019 - 28. Okt. 2019. In: 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW). Piscataway, NJ: IEEE, 2019, pp. 4321-4325. ISSN 2473-9936. eISSN 2473-9944. ISBN 978-1-72815-023-9. Available under: doi: 10.1109/ICCVW.2019.00516

BibTex

@inproceedings{Schlegel2019-10Towar-50801,
  year={2019},
  doi={10.1109/ICCVW.2019.00516},
  title={Towards A Rigorous Evaluation Of XAI Methods On Time Series},
  isbn={978-1-72815-023-9},
  issn={2473-9936},
  publisher={IEEE},
  address={Piscataway, NJ},
  booktitle={2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW)},
  pages={4321--4325},
  author={Schlegel, Udo and Arnout, Hiba and El-Assady, Mennatallah and Oelke, Daniela and Keim, Daniel A.}
}

RDF

<rdf:RDF
    xmlns:dcterms="http://purl.org/dc/terms/"
    xmlns:dc="http://purl.org/dc/elements/1.1/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:bibo="http://purl.org/ontology/bibo/"
    xmlns:dspace="http://digital-repositories.org/ontologies/dspace/0.1.0#"
    xmlns:foaf="http://xmlns.com/foaf/0.1/"
    xmlns:void="http://rdfs.org/ns/void#"
    xmlns:xsd="http://www.w3.org/2001/XMLSchema#" > 
  <rdf:Description rdf:about="https://kops.uni-konstanz.de/server/rdf/resource/123456789/50801">
    <dcterms:title>Towards A Rigorous Evaluation Of XAI Methods On Time Series</dcterms:title>
    <dc:creator>Schlegel, Udo</dc:creator>
    <dcterms:available rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2020-09-11T11:41:53Z</dcterms:available>
    <dc:date rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2020-09-11T11:41:53Z</dc:date>
    <dc:contributor>Keim, Daniel A.</dc:contributor>
    <dc:contributor>El-Assady, Mennatallah</dc:contributor>
    <dspace:hasBitstream rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/50801/1/Schlegel_2-1iaybk1cefn2y7.pdf"/>
    <dc:creator>Keim, Daniel A.</dc:creator>
    <dcterms:issued>2019-10</dcterms:issued>
    <dc:contributor>Oelke, Daniela</dc:contributor>
    <dcterms:rights rdf:resource="https://rightsstatements.org/page/InC/1.0/"/>
    <foaf:homepage rdf:resource="http://localhost:8080/"/>
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dc:language>eng</dc:language>
    <dc:contributor>Arnout, Hiba</dc:contributor>
    <bibo:uri rdf:resource="https://kops.uni-konstanz.de/handle/123456789/50801"/>
    <dc:rights>terms-of-use</dc:rights>
    <dc:creator>Oelke, Daniela</dc:creator>
    <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/>
    <dc:creator>Arnout, Hiba</dc:creator>
    <dc:contributor>Schlegel, Udo</dc:contributor>
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dcterms:hasPart rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/50801/1/Schlegel_2-1iaybk1cefn2y7.pdf"/>
    <dc:creator>El-Assady, Mennatallah</dc:creator>
    <dcterms:abstract xml:lang="eng">Explainable Artificial Intelligence (XAI) methods are typically deployed to explain and debug black-box machine learning models. However, most proposed XAI methods are black-boxes themselves and designed for images. Thus, they rely on visual interpretability to evaluate and prove explanations. In this work, we apply XAI methods previously used in the image and text-domain on time series. We present a methodology to test and evaluate various XAI methods on time series by introducing new verification techniques to incorporate the temporal dimension. We further conduct preliminary experiments to assess the quality of selected XAI method explanations with various verification methods on a range of datasets and inspecting quality metrics on it. We demonstrate that in our initial experiments, SHAP works robust for all models, but others like DeepLIFT, LRP, and Saliency Maps work better with specific architectures.</dcterms:abstract>
  </rdf:Description>
</rdf:RDF>

Universitätsbibliographie

Ja