Automatic interpretation of experimental analytical data of organic molecules
The goal of this project is to use artificial intelligence approaches to mimic human expert knowledge to interpret experimental analytical data. More precisely, we try to build a software which can calculate an organic compound's concentration, evaluate the purity of the compound, and check whether a proposed molecular structure is consistent with the acquired experimental analytical data. The performance of the software should be comparable to the human expert.
To achieve the goal of the project, an expert knowledge base and effective reasoning mechanisms have to be built, which mirror the human interpreter's domain knowledge. Due to the nature of the empiricism and illegibility of human interpretation, we adopt the Bayesian reasoning framework to model the human expert's "knowledge base" and "reasoning mechanisms".