Open bjadidi opened 1 month ago
Hi @bjadidi, I believe most of the training data we have is for organic polymers. I believe the accuracy of the model on organic-inorganic polymer blends would depend on if we have anything similar in the training data to the target, predicted polymers. To evaluate that, we use a "domain of validity" measure (see figure 3A in the polyID paper).
How the model would behave and response to the blends/polymers for organic and inorganic materials?