Non-Robust Systems

This Friday's theory lunch was about using mathematics to quantify the robustness. Basically, the definition of robustness can be considered as a function of input and output.

After back to the lab, we talked about examples in real lives which are robust and non-robust. Just like biology which is very robust, almost every artificial system needs to be robust to functioning well. Thus it is quite easy to find robust systems around us, but probably not for opposed ones.

I thought about one example - the paper peer-reviewing process actually is a very non-robust system. It's pretty unpredictable. Even for the same manuscript (input) sending to the same journal twice, it is not guaranteed the decisions (output) will be always the same. And even worse, this system is very easy to be interfered by outside perturbations. Say, some powerful PIs directly write or call editors arguing the reviews or decisions.

I think it could be one of the best examples in real lives for non-robust systems .