Fig. 2
From: Improving the quality of chemical language model outcomes with atom-in-SMILES tokenization
A toy example illustrating the major differences between AIS and conventional SMILES tokenizations. The formal description of AIS tokenization contains three primary elements, (i) central atom, (ii) ring information, and (iii) neighbor atoms information, interacting with the central atom. The formalism ties everything together within a square bracket separated by a semi-colon. The chirality information can be attached to the central atom, which is labeled with either @ or @ @ suffixes. Aromaticity is reflected on the central atom with a lower case letter. Hydrogen atoms are explicitly specified on central atoms. The hybridization and bonding nature of organic elements can be easily deduced