I'd actually see the main benefit as being stealth and maintainability; i.e, concentrating the hot airflow in the center means that coatings on the exterior, which absorbs traveling waves from the main aircraft body, are subjected either to less heat or less total heat stress. In the former case, you can use less temperature tolerant coatings. In the latter case, maintenance requirements are reduced.Yes, this is a convergent ejector nozzle of the kind that was popular up to the 1960s and is still used by the M88, the RD-33 has a full con/di design (by the looks of it with independent throat/exit control, actually!).
Didn't the WS-10 on the J-11/16 switch to convergent/divergent only recently, well after we started seeing J-20s powered by the engine? Otherwise, one hitherto incidental benefit of the ejector nozzle might make it attractive for use on a LO airframe: the sheath of cool secondary air around the engine exhaust decreases IR signature. It's probably nowhere near as effective a measure for this purpose as a 2D nozzle or something, but it might just be good enough to warrant its retention on the J-20.
As for the IR claim:
It seems that IR emissions vary as:
e^T (with an initial displacing constant). This implies that total IR emissions, depending on other variables might be higher for a small, hot object compared to a larger mass containing the same amount of heat energy.
The localization of the heat energy from the exhaust, of course, makes it easier for the aircraft to hide the high-temperature zone, but that might not be a sufficient trade-off compared to having the hot air from the engine turbine mix with the bypass air.