Figure 3
From: Pareto optimization with small data by learning across common objective spaces
Accuracy of PE measured in \(RMSE_{{\textbf {f}}}\) (y-axis) for different amounts of target training data utilized (x-axis). Results for DTLZ 1a to 3a with 7 objectives (black line) or 4 objectives (grey line) are presented. The marker “o” and “x” represent PS learning by invTGP with single-source transfer and invGP without transfer, respectively.
