- Can an ANN allow us to predict the structure of fluids that are impossible to predict numerically via liquid-state theory?
- Can we learn something about liquid-state theory itself by the nature of the trained ANN?
- What features do the hidden layers capture?
- Can experimental scattering data be used in the training of the ANN?
- How does the mixing of synthetic and experimental data affect predictions?
- Can we use the trained ANN to create a new liquid-state closure?
- Closure approximations account for the majority of error in liquid state theory.