Emergence Attribution Runtime Analysis on Mythos
1.2Wall-clock Runtime (s)Sampled Banzhaf
Evaluation results
| Method | Links | |
|---|---|---|
| Sampled BanzhafN=10^2, Value function=f_heat, T=14, D=3, Hardware=single CPU node, m=1000 samples2026.05 | 1.2 | |
| Exact ShapleyN=10, Value function=f_heat, T=14, D=3, Hardware=single CPU node2026.05 | 1.4 | |
| Exact BanzhafN=10, Value function=f_heat, T=14, D=3, Hardware=single CPU node2026.05 | 1.4 | |
| LOON=10^6, Value function=f_heat, T=14, D=3, Hardware=single CPU node2026.05 | 1.6 | |
| Sampled BanzhafN=10^6, Value function=f_heat, T=14, D=3, Hardware=single CPU node, m=1000 samples2026.05 | 3.8 | |
| Aumann–Shapley path-integral attributionN=10, Value function=f_heat, T=14, D=3, Hardware=single CPU node2026.05 | 4.9 | |
| Aumann–Shapley path-integral attributionN=10^2, Value function=f_heat, T=14, D=3, Hardware=single CPU node2026.05 | 5.1 | |
| LOON=10, Value function=f_heat, T=14, D=3, Hardware=single CPU node2026.05 | 5.7 | |
| LOON=10^2, Value function=f_heat, T=14, D=3, Hardware=single CPU node2026.05 | 5.8 | |
| Sampled ShapleyN=10^2, Value function=f_heat, T=14, D=3, Hardware=single CPU node, m=1000 samples2026.05 | 8.6 | |
| Aumann–Shapley path-integral attributionN=10^6, Value function=f_heat, T=14, D=3, Hardware=single CPU node2026.05 | 8.6 | |
| Sampled ShapleyN=10^6, Value function=f_heat, T=14, D=3, Hardware=single CPU node, m=1000 samples2026.05 | 9.3 | |
| Sampled ShapleyN=10, Value function=f_heat, T=14, D=3, Hardware=single CPU node, m=1000 samples2026.05 | 9.6 | |
| Sampled BanzhafN=10, Value function=f_heat, T=14, D=3, Hardware=single CPU node, m=1000 samples2026.05 | 9.8 |