1. In Step 1, are we restricted to using only the molecular features mentioned in the wiki, or is it permissible to incorporate additional features? 2. Should the model in Step 1 be trained without using SMILES? (i.e., SMILES are used during training but not for test samples because the test data does not include SMILES?) 3. Is it allowed to use additional unlabeled molecules from external sources beyond the provided dataset for training?

Created by Woojae Chang woozch
1. You can use additional features, but as you dont have the molecules identity not sure you can 2. You dont have the SMILES in step 1. 3. Yes you can use any external dataset

Questions Regarding Feature Usage and Data Constraints in Step 1 page is loading…