In this paper the last sentence of the first paragraph in the abstract states:
We assume a (necessary) “leak-free” preprocessing stage.
But then later on in section 1.4 they say
Another interesting open question is to construct a leakage- resilient MPC protocol without assuming any leak-free stages,
Why do they say the leak-free stage is necessary but then later on they say it's an interesting open question ? When they say it's necessary doesn't that imply there is no open question (i.e. not debatable) ?
Is there any research done on leak-free stageS?