ChemRxiv
These are preliminary reports that have not been peer-reviewed. They should not be regarded as conclusive, guide clinical practice/health-related behavior, or be reported in news media as established information. For more information, please see our FAQs.
On failure modes of molecule generators and optimizers_chemrxiv.pdf (2.52 MB)

On Failure Modes of Molecule Generators and Optimizers

preprint
submitted on 29.04.2020 and posted on 30.04.2020 by Philipp Renz, Dries Van Rompaey, Jörg Kurt Wegner, Sepp Hochreiter, Günter Klambauer
There has been a wave of generative models for molecules triggered by advances in the field of Deep Learning. These generative models are often used to optimize chemical compounds towards particular properties or a desired biological activity. The evaluation of generative models remains challenging and suggested performance metrics or scoring functions often do not cover all relevant aspects of drug design projects. In this work, we highlight some unintended failure modes of generative models and how these evade detection by current performance metrics.

Funding

This work was supported by Flanders Innovation & Entrepreneurship (VLAIO) with the project grant HBC.2018.2287 (madeSMART).

History

Email Address of Submitting Author

renz@ml.jku.at

Institution

Institute for Machine Learning, JKU Linz

Country

Austria

ORCID For Submitting Author

0000-0002-3323-7632

Declaration of Conflict of Interest

None declared. JKW and DVR are employed at Janssen Pharmaceutica N.V.

Licence

Exports

Logo branding

Licence

Exports