In manufacturing processes and other situation, modeling of process parameters is important. Which models are better? Artificial Intelligence or statistical models?
Your question would need to be refined. It is too ambiguous.
Some branches of AI uses some non-deterministic techniques to optimise parameters and optimise algorithms to solve a problems. Examples are evolutionary algorithms and others meta-heuristics. (https://en.wikipedia.org/wiki/Evolutionary_algorithm and https://en.wikipedia.org/wiki/Evolutionary_computation). Artificial neural network is another form of nature inspired intelligent agent.
All of these models rely on statistics.
I would interested if you could be more specific and suggests how you would differentiate AI and statistical models.
It is a very good point. Some evolutionary algorithms have a slow convergence, and some others have a much faster one. Hill-climbers (ie the operators can only improve the population over time) has often a much better convergence, but it can lead to a local optima. So the balance is to disrupt enough the solutions so that better solutions can be found.
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling