Pipeline for chemical image-to-text competition
BMS-Molecular-Translation
Pipeline for chemical image-to-text competition.
This is a pipeline for Bristol-Myers Squibb – Molecular Translation by Vadim Timakin and Maksim Zhdanov. We got bronze medals in this competition. Significant part of code was originated from Y.Nakama’s notebook
This competition was about image-to-text translation of images with molecular skeletal strucutures to InChI chemical formula identifiers.
InChI=1S/C16H13Cl2NO3/c1-10-2-4-11(5-3-10)16(21)22-9-15(20)19-14-8-12(17)6-7-13(14)18/h2-8H,9H2,1H3,(H,19,20)
Solution
General Encoder-Decoder concept
Most participants used CNN encoder to acquire features with decoder (LSTM/GRU/Transformer) to get text sequences. That’s a casual approach to image captioning problem.
Pseudo-labelling with InChI validation using RDKit
RDKit is an open source toolkit for cheminformatics and it was quite useful while solving the problem. When we trained our first model, it scored around 7-8