Challenge #1

Collect and standardise chemical reaction data, including structures, reagents, and conditions
- Standardisation of chemical reaction data is challenging because there are varies of reactions in organic chemistry. For example, reagents (such as Pd/C) may have different specifications. Solvents may or may not participate (as a reactant) in the reaction.
- NLP may be required to extract data from reaction procedures.

Challenge #2

Use AI to predict/identify byproducts of reactions with the support of analytical chemistry
- Predict potential byproducts based on reactants (and catalyst etc.).
- Elucidate the spectrum data, such as LCMS and H-NMR with AI （Available MS library: https://gnps.ucsd.edu/ProteoSAFe/static/gnps-splash.jsp）

Challenge #3

Evaluate reactivity/reaction feasibility with quantum mechanics and/or AI technology.
- Predict whether the reaction is feasible based on public reaction data such as Chemical reactions from US patents (1976-Sep2016)
- Collect or generate failure reactions to train and/or evaluate the model
- It is encouraged to use quantum mechanics approaches.