David Pätzel

Wissenschaftlicher Mitarbeiter
Lehrstuhl für Organic Computing
Telefon: +49 821 598 69254
E-Mail:
Raum: 2043 (W)
Sprechzeiten: nach Vereinbarung
Adresse: Am Technologiezentrum 8, 86159 Augsburg

Forschungsschwerpunkte

My main research is directed towards metaheuristic Rule Set Learning algorithms (MRSLs) for regression problems. MRSLs are machine learning algorithms that build sets of if-then rules using metaheuristics such as evolutionary algorithms. One of the main arguments for applying MRSLs is that they generate models that are more easily interpreted than the models generated by other learning systems such as neural networks.

While there are several such algorithms (for example, XCSF from the MRSL subgroup of Learning Classifier Systems), it is not well-understood how they work or even when they work well. Even more so, it is not even clear whether the metaheuristics (and their operators) that they build upon are anywhere near optimal for the task of searching rule set space.

With my current my work I'm looking into exactly that: For one, I'm developing novel ways to benchmark these systems such that fresh insights into rule-set-searching metaheuristics can be gained. Further, I'm looking into applying metaheuristic operators to these tasks that have not yet been applied to them.

Other than that, I'm very interested in

  • evolutionary machine learning algorithms
  • reinforcement learning
  • functional programming, especially using Haskell
  • ways to incoporate a more formal (functional) view into everyday software such as operating systems (NixOS) or window managers (XMonad)
  • free software

Sonstige wissenschaftliche Tätigkeiten

Ich habe Reviews für Artikel folgender Journals verfasst: Daneben war ich Teil des gewählten dreiköpfigen Organisationskomitees des 23., 24., 25. und 26. International Workshop Evolutionary Rule-based Machine Learning (ERBML) (ehemals International Workshop on Learning Classifier Systems, IWLCS), der im Rahmen der Genetic and Evolutionary Computation Conference (GECCO) in den Jahren 2020 bis 2023 stattfand. Seitdem bin ich auch Reviewer für auf dem ERBML eingereichte Papiere.

Lebenslauf

seit 2017 Wissenschaftlicher Mitarbeiter am Lehrstuhl
2015–2017 Master-Studium im Fach Informatik an der Universität Augsburg
2011–2015 Bachelor-Studium im Fach Informatik und Multimedia an der Universität Augsburg

Lehrveranstaltungen

(Angewandte Filter: Semester: aktuelles | Dozenten: David Pätzel | Vorlesungsarten: Übung, Seminar)
Name Semester Typ
Kommunikationssysteme - Übungsgruppe "Marconi" (Mi 14:00) Wintersemester 2023/24 Übung
Übung zu Kommunikationssysteme Wintersemester 2023/24 Übung
Kommunikationssysteme - Übungsgruppe "Floyd" (Di 10:00) Wintersemester 2023/24 Übung
Seminar Organic Computing (Bachelor) Wintersemester 2023/24 Seminar
Übung zu Grundlagen des Organic Computing Wintersemester 2023/24 Übung
Seminar Organic Computing (Master) Wintersemester 2023/24 Seminar
Kommunikationssysteme - Übungsgruppe "Bell" (Mo 15:45) Wintersemester 2023/24 Übung
Kommunikationssysteme - Übungsgruppe "Metcalfe" (Mi 15:45) Wintersemester 2023/24 Übung

Publikationen

2023 | 2022 | 2021 | 2020 | 2019 | 2018

2023

Michael Heider, David Pätzel, Helena Stegherr and Jörg Hähner. 2023. A metaheuristic perspective on learning classifier systems. DOI: 10.1007/978-981-19-3888-7_3
BibTeX | RIS | DOI

Michael Heider, Helena Stegherr, David Pätzel, Roman Sraj, Jonathan Wurth, Benedikt Volger and Jörg Hähner. 2023. Discovering rules for rule-based machine learning with the help of novelty search. DOI: 10.1007/s42979-023-02198-x
PDF | BibTeX | RIS | DOI

Michael Heider, Helena Stegherr, Roman Sraj, David Pätzel, Jonathan Wurth and Jörg Hähner. 2023. SupRB in the context of rule-based machine learning methods: a comparative study. DOI: 10.1016/j.asoc.2023.110706
BibTeX | RIS | DOI

David Pätzel, Michael Heider and Jörg Hähner. 2023. Towards principled synthetic benchmarks for explainable rule set learning algorithms. DOI: 10.1145/3583133.3596416
BibTeX | RIS | DOI

Henning Cui, David Pätzel, Andreas Margraf and Jörg Hähner. 2023. Weighted mutation of connections to mitigate search space limitations in Cartesian Genetic Programming. DOI: 10.1145/3594805.3607130
PDF | BibTeX | RIS | DOI

2022

Lukas Rosenbauer, David Pätzel, Anthony Stein and Jörg Hähner. 2022. A learning classifier system for automated test case prioritization and selection. DOI: 10.1007/s42979-022-01255-1
BibTeX | RIS | DOI

Michael Heider, David Pätzel and Alexander R. M. Wagner. 2022. An overview of LCS research from 2021 to 2022. DOI: 10.1145/3520304.3533985
PDF | BibTeX | RIS | DOI

Michael Heider, Helena Stegherr, David Pätzel, Roman Sraj, Jonathan Wurth, Benedikt Volger and Jörg Hähner. 2022. Approaches for rule discovery in a learning classifier system. DOI: 10.5220/0011542000003332
PDF | BibTeX | RIS | DOI

David Pätzel and Jörg Hähner. 2022. The Bayesian learning classifier system: implementation, replicability, comparison with XCSF. DOI: 10.1145/3512290.3528736
BibTeX | RIS | DOI

2021

Lukas Rosenbauer, David Pätzel, Anthony Stein and Jörg Hähner. 2021. An organic computing system for automated testing. DOI: 10.1007/978-3-030-81682-7_9
BibTeX | RIS | DOI

David Pätzel, Michael Heider and Alexander R. M. Wagner. 2021. An overview of LCS research from 2020 to 2021. DOI: 10.1145/3449726.3463173
PDF | BibTeX | RIS | DOI

Lukas Rosenbauer, David Pätzel, Anthony Stein and Jörg Hähner. 2021. Transfer learning for automated test case prioritization using XCSF. DOI: 10.1007/978-3-030-72699-7_43
BibTeX | RIS | DOI | URL

2020

David Pätzel, Anthony Stein and Masaya Nakata. 2020. An overview of LCS research from IWLCS 2019 to 2020. DOI: 10.1145/3377929.3398105
BibTeX | RIS | DOI

Michael Heider, David Pätzel and Jörg Hähner. 2020. SupRB: a supervised rule-based learning system for continuous problems.
BibTeX | RIS | URL

Michael Heider, David Pätzel and Jörg Hähner. 2020. Towards a Pittsburgh-style LCS for learning manufacturing machinery parametrizations. DOI: 10.1145/3377929.3389963
PDF | BibTeX | RIS | DOI

Lukas Rosenbauer, Anthony Stein, Roland Maier, David Pätzel and Jörg Hähner. 2020. XCS as a reinforcement learning approach to automatic test case prioritization. DOI: 10.1145/3377929.3398128
BibTeX | RIS | DOI

Lukas Rosenbauer, Anthony Stein, David Pätzel and Jörg Hähner. 2020. XCSF for automatic test case prioritization. DOI: 10.5220/0010105700490058
PDF | BibTeX | RIS | DOI

Lukas Rosenbauer, Anthony Stein, David Pätzel and Jörg Hähner. 2020. XCSF with experience replay for automatic test case prioritization. DOI: 10.1109/ssci47803.2020.9308379
BibTeX | RIS | DOI

2019

David Pätzel, Anthony Stein and Jörg Hähner. 2019. A survey of formal theoretical advances regarding XCS. DOI: 10.1145/3319619.3326848
BibTeX | RIS | DOI

2018

David Pätzel and Jörg Hähner. 2018. An algebraic description of XCS. DOI: 10.1145/3205651.3208248
BibTeX | RIS | DOI

Suche