English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Computer-based de-novo design of functional molecules is one of the mostprominent challenges in cheminformatics today. As a result, generative andevolutionary inverse designs from the field of artificial intelligence haveemerged at a rapid pace, with aims to optimize molecules for a particularchemical property. These models 'indirectly' explore the chemical space; bylearning latent spaces, policies, distributions or by applying mutations onpopulations of molecules. However, the recent development of the SELFIES stringrepresentation of molecules, a surjective alternative to SMILES, have madepossible other potential techniques. Based on SELFIES, we therefore proposePASITHEA, a direct gradient-based molecule optimization that appliesinceptionism techniques from computer vision. PASITHEA exploits the use ofgradients by directly reversing the learning process of a neural network, whichis trained to predict real-valued chemical properties. Effectively, this formsan inverse regression model, which is capable of generating molecular variantsoptimized for a certain property. Although our results are preliminary, weobserve a shift in distribution of a chosen property during inverse-training, aclear indication of PASITHEA's viability. A striking property of inceptionismis that we can directly probe the model's understanding of the chemical spaceit was trained on. We expect that extending PASITHEA to larger datasets,molecules and more complex properties will lead to advances in the design ofnew functional molecules as well as the interpretation and explanation ofmachine learning models.

Deep Molecular Dreaming: Inverse machine learning for de-novo molecular design and interpretability with surjective representations