Evolutionary techniques in AutoML
Thesis title in Czech: | Evoluční techniky v automatickém strojovém učení |
---|---|
Thesis title in English: | Evolutionary techniques in AutoML |
Key words: | Strojové učení|AutoML|evoluční algoritmy |
English key words: | Machine learning|AutoML|evolutionary algorithms |
Academic year of topic announcement: | 2020/2021 |
Thesis type: | diploma thesis |
Thesis language: | angličtina |
Department: | Department of Theoretical Computer Science and Mathematical Logic (32-KTIML) |
Supervisor: | Mgr. Roman Neruda, CSc. |
Author: | Rajat Sharma - assigned and confirmed by the Study Dept. |
Date of registration: | 14.09.2021 |
Date of assignment: | 14.09.2021 |
Confirmed by Study dept. on: | 26.04.2022 |
Guidelines |
AutoML methods search for a suitable pipeline of preprocessing and ML components given a data set representing an ML task. The goal of the thesis is to design an evolutionary optimization algorithm which will search the space of pipelines and propose an optimized solution. Student will design a workflow representation, as well as other components of the evolutionary algorithm. Several approaches, such as cartesian genetic programming, or grammar based evolution will be tested. The approach should deal with optimal hyper-parameter selection for used methods. Implementation of developed algorithms using standard machine learning libraries such as scikit-learn, and their experimental evaluation on benchmark data will be a part of the work. |
References |
[1] Riccardo Poli, William B. Langdon, Nicholas Freitag McPhee: A field guide to genetic programming, Published by Lulu.com, http://www.gp-field-guide.org.uk. (2008)
[2] Peter Flach: Machine Learning: The Art and Science of Algorithms That Make Sense of Data, Cambridge University Press. (2012) [3] Pavel Brazdil, Christophe Giraud-Carrier, Carlos Soares, Ricardo Vilalta: Metalearning: Applications to Data Mining. Springer. (2008) [4] Miller, Julian F., ed. (2011). Cartesian Genetic Programming, Springer 2011 [5] Herausgeber: Hutter, Frank, Kotthoff, Lars, Vanschoren, Joaquin (Eds.). Automated MachineL earning: Methods, Systems, Challenges, Springer, 2019 |