0.6 0.5 0.4 0.3 0.2 0.1 0.0 Training Epoch: 50 20 50 100 Search Budget (a) 0.6 0.5 0.4 0.3 0.2 0.1 0.0 20 Training Epoch: 150 0.6 0.5 0.4 0.3 0.2 0.1 0.0 50 100 Search Budget (c) 0.6 0.5 0.4 0.3 0.2 0.1 0.0 20 Training Epoch: 250 0.6 0.5 0.4 0.3 0.2 0.1 0.0 50 100 Search Budget (e) 150 200 20 150 200 20 150 200 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Training Epoch: 100 20 50 100 Search Budget (b) Training Epoch: 200 150 200 50 100 Search Budget (d) Training Epoch: 300 150 200 50 100 Search Budget (f) SUPERVISED SS-CCL-10K SS-CCL-40K SS-CCL-70K SS-CCL-100K SS-CCL-140K 150 200 FIGURE 6 Predictive performance of neural predictors pre-trained by different batch size N on NASBench-101. AUGUST 2021 | IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE 45 Kendall Tau Correlation Kendall Tau Correlation Kendall Tau Correlation Kendall Tau Correlation Kendall Tau Correlation Kendall Tau Correlation