1. Generalization ability of feature-based performance prediction models : a statistical analysis across benchmarksAna Nikolikj, Ana Kostovska, Gjorgjina Cenikj, Carola Doerr, Tome Eftimov, 2024, objavljeni znanstveni prispevek na konferenci Povzetek: This study examines the generalization ability of algorithm performance prediction models across various bench-mark suites. Comparing the statistical similarity between the problem collections with the accuracy of performance prediction models that are based on exploratory landscape analysis features, we observe that there is a positive correlation between these two measures. Specifically, when the high-dimensional feature value distributions between training and testing suites lack statistical significance, the model tends to generalize well, in the sense that the testing errors are in the same range as the training errors. Two experiments validate these findings: one involving the standard benchmark suites, the BBOB and CEC collections, and another using five collections of affine combinations of BBOB problem instances.
Ključne besede: meta-learning, single-objective optimization, module importance
Objavljeno v DiRROS: 16.09.2024; Ogledov: 41; Prenosov: 25
 Celotno besedilo (1,29 MB)
Gradivo ima več datotek! Več... |
2. |
3. PS-AAS : portfolio selection for automated algorithm selection in black-box optimizationAna Kostovska, Gjorgjina Cenikj, Diederick Vermetten, Anja Janković, Ana Nikolikj, Urban Škvorc, Peter Korošec, Carola Doerr, Tome Eftimov, 2023, objavljeni znanstveni prispevek na konferenci Ključne besede: automated algorithm selection, portfolio selection, black box optimization
Objavljeno v DiRROS: 11.12.2023; Ogledov: 643; Prenosov: 260
Celotno besedilo (1,90 MB)
Gradivo ima več datotek! Več... |
4. Sensitivity analysis of RF+clust for leave-one-problem-out performance predictionAna Nikolikj, Michal Pluhacek, Carola Doerr, Peter Korošec, Tome Eftimov, 2023, objavljeni znanstveni prispevek na konferenci Ključne besede: automated performance prediction, autoML, single-objective black-box optimization, zero-shot learning
Objavljeno v DiRROS: 13.11.2023; Ogledov: 610; Prenosov: 391
Celotno besedilo (4,94 MB)
Gradivo ima več datotek! Več... |
5. Algorithm instance footprint : separating easily solvable and challenging problem instancesAna Nikolikj, Sašo Džeroski, Mario Andrés Muñoz, Carola Doerr, Peter Korošec, Tome Eftimov, 2023, objavljeni znanstveni prispevek na konferenci Ključne besede: black-box optimization, algorithms, problem instances, machine learning
Objavljeno v DiRROS: 15.09.2023; Ogledov: 545; Prenosov: 318
Celotno besedilo (2,03 MB)
Gradivo ima več datotek! Več... |
6. Assessing the generalizability of a performance predictive modelAna Nikolikj, Gjorgjina Cenikj, Gordana Ispirova, Diederick Vermetten, Ryan Dieter Lang, Andries Petrus Engelbrecht, Carola Doerr, Peter Korošec, Tome Eftimov, 2023, objavljeni znanstveni prispevek na konferenci Ključne besede: algorithms, predictive models, machine learning
Objavljeno v DiRROS: 15.09.2023; Ogledov: 604; Prenosov: 380
Celotno besedilo (935,67 KB)
Gradivo ima več datotek! Več... |
7. |
8. DynamoRep : trajectory-based population dynamics for classification of black-box optimization problemsGjorgjina Cenikj, Gašper Petelin, Carola Doerr, Peter Korošec, Tome Eftimov, 2023, objavljeni znanstveni prispevek na konferenci Ključne besede: black-box single-objective optimization, optimization problem classification, problem representation, meta-learning
Objavljeno v DiRROS: 30.08.2023; Ogledov: 597; Prenosov: 387
Celotno besedilo (650,13 KB)
Gradivo ima več datotek! Več... |
