QSAR (Quantitative Structure-Activity Relationship) uses neural networks to determine the effects of an input molecule on the human body. It has been very successful and is attributed with helping find cancer drugs and even the COVID vaccine. Some 20 years ago mQSAR was introduced (multi-dimensional QSAR). But due to complex parameters and other issues it failed to produce results. The parametric approach to AI didn't seem to suffice I've come up with a new approach to ML and have been posting on it in these forums. I think this can support mQSAR from preliminary work. I have contacted William Goddard, a great genius in computational chemistry and co-founder of Schrodinger, a company in NY that manufacturers QSAR systems. They may want to look to support mQSAR, it could be an opportunity for IBM.
------------------------------
John Harby
CEO
Autonomic AI, LLC
Temecula CA
9513835000
------------------------------