Does the cross-scale k_s work in the way described in the paper?

[Leopold2333]

CrossGNN/models/CrossGNN.py

Line 161 in 0407abd

min_total_corss_scale_neighbors = 5 # number

CrossGNN/models/CrossGNN.py

Line 171 in 0407abd

kp=max(kp,min_total_corss_scale_neighbors)

Hi, I'm confused of this min_total_corss_scale_neighbors variable here. If there is 'min_total_corss_scale_neighbors', then what's the meaning of the existence of $k_s$? The paper does not mention about the variable, is there any reason why it would improve performance? If so, it feels like that $k_s$ doesn't consistently act as a constraint.