Finding the DeepDream for Time Series: Activation Maximization for Univariate Time Series

Understanding how models process and interpret time series data remains a significant challenge in deep learning to enable applicability in safety-critical areas such as healthcare. In this paper, we introduce Sequence Dreaming, a technique that adapts Activation Maximization to analyze sequential i...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Schlegel, Udo, Keim, Daniel A, Sutter, Tobias
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Schlegel, Udo
Keim, Daniel A
Sutter, Tobias
description Understanding how models process and interpret time series data remains a significant challenge in deep learning to enable applicability in safety-critical areas such as healthcare. In this paper, we introduce Sequence Dreaming, a technique that adapts Activation Maximization to analyze sequential information, aiming to enhance the interpretability of neural networks operating on univariate time series. By leveraging this method, we visualize the temporal dynamics and patterns most influential in model decision-making processes. To counteract the generation of unrealistic or excessively noisy sequences, we enhance Sequence Dreaming with a range of regularization techniques, including exponential smoothing. This approach ensures the production of sequences that more accurately reflect the critical features identified by the neural network. Our approach is tested on a time series classification dataset encompassing applications in predictive maintenance. The results show that our proposed Sequence Dreaming approach demonstrates targeted activation maximization for different use cases so that either centered class or border activation maximization can be generated. The results underscore the versatility of Sequence Dreaming in uncovering salient temporal features learned by neural networks, thereby advancing model transparency and trustworthiness in decision-critical domains.
doi_str_mv 10.48550/arxiv.2408.10628
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2408_10628</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2408_10628</sourcerecordid><originalsourceid>FETCH-arxiv_primary_2408_106283</originalsourceid><addsrcrecordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjGw0DM0MDOy4GSIdMvMS8nMS1coyUhVcElNLXApSk3MVUjLL1IIycxNVQhOLcpMLbZScEwuySxLLMnMz1PwTazIzM2sgnBACkPzgFJFmYklqch6eBhY0xJzilN5oTQ3g7yba4izhy7YEfEFRZm5iUWV8SDHxIMdY0xYBQArfD-R</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Finding the DeepDream for Time Series: Activation Maximization for Univariate Time Series</title><source>arXiv.org</source><creator>Schlegel, Udo ; Keim, Daniel A ; Sutter, Tobias</creator><creatorcontrib>Schlegel, Udo ; Keim, Daniel A ; Sutter, Tobias</creatorcontrib><description>Understanding how models process and interpret time series data remains a significant challenge in deep learning to enable applicability in safety-critical areas such as healthcare. In this paper, we introduce Sequence Dreaming, a technique that adapts Activation Maximization to analyze sequential information, aiming to enhance the interpretability of neural networks operating on univariate time series. By leveraging this method, we visualize the temporal dynamics and patterns most influential in model decision-making processes. To counteract the generation of unrealistic or excessively noisy sequences, we enhance Sequence Dreaming with a range of regularization techniques, including exponential smoothing. This approach ensures the production of sequences that more accurately reflect the critical features identified by the neural network. Our approach is tested on a time series classification dataset encompassing applications in predictive maintenance. The results show that our proposed Sequence Dreaming approach demonstrates targeted activation maximization for different use cases so that either centered class or border activation maximization can be generated. The results underscore the versatility of Sequence Dreaming in uncovering salient temporal features learned by neural networks, thereby advancing model transparency and trustworthiness in decision-critical domains.</description><identifier>DOI: 10.48550/arxiv.2408.10628</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Learning</subject><creationdate>2024-08</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2408.10628$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2408.10628$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Schlegel, Udo</creatorcontrib><creatorcontrib>Keim, Daniel A</creatorcontrib><creatorcontrib>Sutter, Tobias</creatorcontrib><title>Finding the DeepDream for Time Series: Activation Maximization for Univariate Time Series</title><description>Understanding how models process and interpret time series data remains a significant challenge in deep learning to enable applicability in safety-critical areas such as healthcare. In this paper, we introduce Sequence Dreaming, a technique that adapts Activation Maximization to analyze sequential information, aiming to enhance the interpretability of neural networks operating on univariate time series. By leveraging this method, we visualize the temporal dynamics and patterns most influential in model decision-making processes. To counteract the generation of unrealistic or excessively noisy sequences, we enhance Sequence Dreaming with a range of regularization techniques, including exponential smoothing. This approach ensures the production of sequences that more accurately reflect the critical features identified by the neural network. Our approach is tested on a time series classification dataset encompassing applications in predictive maintenance. The results show that our proposed Sequence Dreaming approach demonstrates targeted activation maximization for different use cases so that either centered class or border activation maximization can be generated. The results underscore the versatility of Sequence Dreaming in uncovering salient temporal features learned by neural networks, thereby advancing model transparency and trustworthiness in decision-critical domains.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjGw0DM0MDOy4GSIdMvMS8nMS1coyUhVcElNLXApSk3MVUjLL1IIycxNVQhOLcpMLbZScEwuySxLLMnMz1PwTazIzM2sgnBACkPzgFJFmYklqch6eBhY0xJzilN5oTQ3g7yba4izhy7YEfEFRZm5iUWV8SDHxIMdY0xYBQArfD-R</recordid><startdate>20240820</startdate><enddate>20240820</enddate><creator>Schlegel, Udo</creator><creator>Keim, Daniel A</creator><creator>Sutter, Tobias</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240820</creationdate><title>Finding the DeepDream for Time Series: Activation Maximization for Univariate Time Series</title><author>Schlegel, Udo ; Keim, Daniel A ; Sutter, Tobias</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-arxiv_primary_2408_106283</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Schlegel, Udo</creatorcontrib><creatorcontrib>Keim, Daniel A</creatorcontrib><creatorcontrib>Sutter, Tobias</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Schlegel, Udo</au><au>Keim, Daniel A</au><au>Sutter, Tobias</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Finding the DeepDream for Time Series: Activation Maximization for Univariate Time Series</atitle><date>2024-08-20</date><risdate>2024</risdate><abstract>Understanding how models process and interpret time series data remains a significant challenge in deep learning to enable applicability in safety-critical areas such as healthcare. In this paper, we introduce Sequence Dreaming, a technique that adapts Activation Maximization to analyze sequential information, aiming to enhance the interpretability of neural networks operating on univariate time series. By leveraging this method, we visualize the temporal dynamics and patterns most influential in model decision-making processes. To counteract the generation of unrealistic or excessively noisy sequences, we enhance Sequence Dreaming with a range of regularization techniques, including exponential smoothing. This approach ensures the production of sequences that more accurately reflect the critical features identified by the neural network. Our approach is tested on a time series classification dataset encompassing applications in predictive maintenance. The results show that our proposed Sequence Dreaming approach demonstrates targeted activation maximization for different use cases so that either centered class or border activation maximization can be generated. The results underscore the versatility of Sequence Dreaming in uncovering salient temporal features learned by neural networks, thereby advancing model transparency and trustworthiness in decision-critical domains.</abstract><doi>10.48550/arxiv.2408.10628</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2408.10628
ispartof
issn
language eng
recordid cdi_arxiv_primary_2408_10628
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Learning
title Finding the DeepDream for Time Series: Activation Maximization for Univariate Time Series
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-10T13%3A31%3A24IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Finding%20the%20DeepDream%20for%20Time%20Series:%20Activation%20Maximization%20for%20Univariate%20Time%20Series&rft.au=Schlegel,%20Udo&rft.date=2024-08-20&rft_id=info:doi/10.48550/arxiv.2408.10628&rft_dat=%3Carxiv_GOX%3E2408_10628%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true