You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If you want to get several molecules similar to a given seed molecule m, you can encode it, perturb its embedding by random noise, and then decode back. If the noise is too small, you'll get many copies of m back, while if it's too large, the molecules won't be that similar; if you tune the noise magnitude just right you'll get "neighbors" of m. You can see Appendix F in the arxiv version of the paper for more details.
No description provided.
The text was updated successfully, but these errors were encountered: