Exploring the Impact of Data Scale on Mutation Step Size in SLIM-GSGP

Farinati, Davide; Pietropolli, Gloria; Vanneschi, Leonardo

http://hdl.handle.net/10362/182852

Utilize este identificador para referenciar este registo.

Nome:	Descrição:	Tamanho:	Formato:
Impact_of_Data_Scale_on_Mutation_Step_Size_SLIM-GSGP_AAM.pdf		1.87 MB	Adobe PDF	Ver/Abrir

Contacte-nos

Autores

Farinati, Davide

Pietropolli, Gloria

Vanneschi, Leonardo

Resumo(s)

The Semantic Learning algorithm based on Inflate and deflate Mutation (SLIM) is a promising recent variant of Geometric Semantic Genetic Programming (GSGP) that introduces a new Deflate Geometric Semantic Mutation (DGSM). This operator maintains the key feature of the standard Geometric Semantic Mutation (GSM), inducing a unimodal error surface for any supervised learning problem, while generating smaller offspring than their parents, and thus allowing SLIM to generate compact, and potentially interpretable, final solutions. A key parameter controlling the evolution process in both GSGP and SLIM is the Mutation Step (MS), which regulates the extent of perturbation to the parent semantics. While it is intuitive that the optimal value of MS has a relationship with the scale of the dataset features, to the best of our knowledge no prior research has extensively explored this relationship. In this work, we provide the first comprehensive investigation into this topic. First, we hypothesize a general rule by analyzing results from artificial datasets, and then we confirm these findings with more complex, real-world datasets. This approach offers a solid alternative to the typical hyperparameter tuning approach.

Descrição

Farinati, D., Pietropolli, G., & Vanneschi, L. (2025). Exploring the Impact of Data Scale on Mutation Step Size in SLIM-GSGP. In B. Xue, L. Manzoni, & I. Bakurov (Eds.), Genetic Programming: 28th European Conference, EuroGP 2025, Held as Part of EvoStar 2025, Trieste, Italy, April 23–25, 2025, Proceedings (pp. 35-51). (Lecture Notes in Computer Science; Vol. 15609). Springer Nature Switzerland AG. https://doi.org/10.1007/978-3-031-89991-1_3 --- This work was supported by national funds through FCT (Fundação para a Ciência e a Tecnologia), under the project - UIDB/04152/2020 - Centro de Investigação em Gestão de Informação (MagIC)/NOVA IMS (https://doi.org/10.54499/UIDB/04152/2020).

Palavras-chave

Genetic Programming Geometric Semantic Genetic Programming Geometric Mutation Mutation Step Symbolic Regression Theoretical Computer Science General Computer Science