The goal of this workshop is to bring together the main scientific players who are likely to deepen our understanding of chemical compound space (CCS) using quantum mechanics (QM) and machine learning (ML). Participants will strongly benefit from the mutual exchange of ideas in this upcoming field of science. The specific scientific domains include statistical mechanics, liquid and solid state physics, quantum chemistry, graph theory, molecular physics, condensed matter physics, optimization algorithms, data mining, and statistical learning.
One can formulate the specific ideas and topics, to be discussed and tackled, in the form of questions.
-How can the CCS be searched most effectively to find a compound with a target property?
-Do Taylor expansions in CCS converge, and to which order?
-Can higher order derivatives in CCS be approximated efficiently?
-How large are errors introduced due to the use of approximate exchange-correlation potentials?
-Is there a rigorous QM derived definition of transferability?
-Is there a rigorous ML derived definition of representative molecular sets?
-What is the nature of the interplay between similarity measures in ML models and training set composition?
-Is there a rigorous relationship between a ML model universal in CCS and QM (which is also universal in CCS)?
-There are many ways to measure distance (or similarity) in CCS, is there an optimal measure, possibly derived from QM?
-What is the exact relationship between alchemical Taylor expansions and ML models?