The aim is to learn physicochemical properties of interest from simple structural representations, using deep neural networks: in particular I would like to investigate how much can be learned from molecular fingerprints generated by using the SMILES molecular representation.
![](https://www.researchgate.net/profile/Dong-Sheng_Cao/publication/235919348/figure/fig1/AS:299748008448007@1448476902809/Figure-1-Representation-of-a-molecular-substructure-fingerprint-with-a-substructure.png)
This project was initiated in 0ctober 2017 during the IPAM’s long program “Complex High-Dimensional Energy Landscapes”