Searcharxiv⌕ Search

arXiv subjects

Toshiharu Hatanaka

Publications and source records attributed to Toshiharu Hatanaka.

2 recordsLinked to original sources

Shortcut Learning in Feature-Based Algorithm Selection

Landscape features give learning models information about the objective function for algorithm selection, algorithm configuration, and performance prediction. Should these features capture more information? Not always, because a model may rely on information that the algorithms ignore. This phenomenon is called shortcut learning, and in-distribution accuracy can hide it. We propose the Rubber Ruler Test. It transforms the objective values of test instances with random strictly increasing transformations, reruns the algorithms, and measures prediction performance. Comparison-based algorithms such as CMA-ES perform identically after the transformation, so for them any drop reveals shortcut learning. On 10-dimensional BBOB, a model for two comparison-based algorithms learned shortcuts even from normalized objective values. The R^2 of a model using all ELA features dropped from 0.85 to 0.23, below 0.45 with invariant features only. Rank-based features reduced the drop when algorithms search using ranks. Data augmentation with transformed training instances improved prediction on transformed instances for every portfolio we studied, but lowered R^2 on original instances by up to 0.31. These results suggest that models should rely only on information that the algorithms use. We recommend reporting the Rubber Ruler Test alongside in-distribution accuracy and using it to choose feature extraction and data augmentation.

cs.NE↗

Novelty-Based Generation of Continuous Landscapes with Diverse Local Optima Networks

Local Optima Networks (LONs) represent the global structure of search spaces as graphs, but their construction requires iterative execution of a search algorithm to find local optima and approximate transitions between Basins of Attraction (BoAs). In continuous optimization, this high computational cost prevents systematic investigation of the relationship between LON features and evolutionary algorithm performance. To address this issue, we propose an alternative definition of BoAs for Max-Set of Gaussians (MSG) landscapes with explicitly tunable multimodality. This bypasses search-based BoA identification, enabling low-cost LON construction. Moreover, we leverage Novelty Search (NS) to explore the parameter space of the MSG landscape generator, producing instances with diverse graph topologies. Our experiments show that the proposed BoAs closely align with gradient-based BoAs, and that NS successfully generates instances with varied search difficulty and connectivity patterns among optima. Finally, over the instances generated by NS, we predict the success rate of two well-established evolutionary algorithms from LON features. While our LON construction is specific to MSG landscapes, the proposed framework provides a dataset that serves as a foundation for landscape-aware optimization.

cs.NE↗