Publication ● 27 Jun 2020Molecule Edit Graph Attention Network: Modeling Chemical Reactions as Sequences of Graph Edits We propose Molecule Edit Graph Attention Network (MEGAN), a template-free neural model that encodes reaction as a sequence of graph edit.
Publication ● 20 Jun 2020We Should at Least Be Able to Design Molecules That Dock Well We propose a benchmark based on docking, a popular computational method for assessing molecule binding to a protein.
Our technology is approximately two orders of magnitude faster than most competitors. This enables us to test significantly more approaches and build otherwise unachievable features.
We have achieved state-of-the-art performance on a standard benchmark for retrosynthesis: better than all published approaches, including models used by top competitors.
Our models do not only give accurate predictions, but also explain them to chemists and give relevant estimates of uncertainty. This is crucial for execution of the proposed plans in the laboratory.
Most competing solutions were designed 15+ years ago and do not utilise recent advances in UX design. This is key to achieve wide adoption relatively quickly.
There is no “one-size-fits-all” when it comes to synthesis planning. We offer our clients many customisations, ranging from adapting search criteria to rapid API development for internal integrations.
ble ble ble The “black box” nature of AI makes it challenging to understand why certain predictions are made. Our focus on model interpretability makes us well -positioned to collect more meaningful data faster.
Contact us by sending email to [email protected] or use form below
Public beta launch
Full product launch
To book one contact us by email or form below
Accoun setup within minutes
Leverage the power of AI into your business
COVID Molecule.one updates ● 24 Mar 2020Molecule.one releases free access to synthesis planning platform to help researchers fight COVID-19
Molecule.one updates ● 02 Oct 2019Molecule.one is a TC Battlefield Startup! We're just about minutes before we start pitching at the TechCrunch Startup Battlefield contest!
Contact us today to find out more about what we do and how we do it. Prototype testing available on request.