# Frugal Algorithm Selection

arXiv: https://arxiv.org/abs/2405.11059 · HTML: https://mlsanity.com/paper/2405.11059/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 49 references that are in the map.
- First posted on arXiv 17 May 2024.

## Builds on (most cited first, 8 of 8)

- [Automated Algorithm Selection: Survey and Perspectives](https://mlsanity.com/paper/1811.11597/) (2018)
- [ASlib: A Benchmark Library for Algorithm Selection](https://mlsanity.com/paper/1506.02465/) (2015)
- [modAL: A modular active learning framework for Python](https://mlsanity.com/paper/1805.00979/) (2018)
- [SELECTOR: Selecting a Representative Benchmark Suite for Reproducible Statistical Comparison](https://mlsanity.com/paper/2204.11527/) (2022)
- [claspfolio 2: Advances in Algorithm Selection for Answer Set Programming](https://mlsanity.com/paper/1405.1520/) (2014)
- [sunny-as2: Enhancing SUNNY for Algorithm Selection](https://mlsanity.com/paper/2009.03107/) (2020)
- [A Framework for Generating Informative Benchmark Instances](https://mlsanity.com/paper/2205.14753/) (2022)
- [Active learning to optimise time-expensive algorithm selection](https://mlsanity.com/paper/1909.03261/) (2019)

## Abstract (arXiv)

When solving decision and optimisation problems, many competing algorithms (model and solver choices) have complementary strengths. Typically, there is no single algorithm that works well for all instances of a problem. Automated algorithm selection has been shown to work very well for choosing a suitable algorithm for a given instance. However, the cost of training can be prohibitively large due to running candidate algorithms on a representative set of training instances. In this work, we explore reducing this cost by choosing a subset of the training instances on which to train. We approach this problem in three ways: using active learning to decide based on prediction uncertainty, augmenting the algorithm predictors with a timeout predictor, and collecting training data using a progressively increasing timeout. We evaluate combinations of these approaches on six datasets from ASLib and present the reduction in labelling cost achieved by each option.