Quantifying Efficiency in Quality Diversity Optimization
Workshop on Benchmarks for Quality-Diversity Algorithms
GECCO 2022 — 10 July 2022
Bryon Tjanaka
University of Southern California
tjanaka@usc.edu
Matthew C. Fontaine
University of Southern California
mfontain@usc.edu
Stefanos Nikolaidis
University of Southern California
nikolaid@usc.edu
Measuring Performance in QD
$$\text{QD score} = \sum_{i=1}^M f(\bm{\phi}_i)$$
Greater performance? No problem!
Similar performance?
Similar performance? AUC!
QD score AUC
- Sfikas et al. 2021 (Monte Carlo Elites)
- Tjanaka et al. 2022 (DQD-RL)
$$\text{QD score AUC} = \sum_{i=1}^N (\lambda * \text{QD score at iteration $i$})$$
Limitations
Further Evaluations
"Intersections"
Applications
- Similar final performance
- Sample efficiency
Quantifying Efficiency in Quality Diversity Optimization
Workshop on Benchmarks for Quality-Diversity Algorithms
GECCO 2022 — 10 July 2022
Bryon Tjanaka, Matthew C. Fontaine, Stefanos Nikolaidis
