I have started my 2 months visiting period at <b>Princeton University<b>!

My research visiting period in Princeton University took place in the group of Professor H. Vincent Poor. The focus of my stay was to work on neural methods for channel capacity estimation using $f$-divergence. Estimating channel capacity with $f$-divergence is particularly complex as there is no general guarantee that maximizing the $f$-MI leads to the capacity-achieving input distribution. We provide and analyze 3 different classes of objective functions, and design a set of self-consistency tests to verify the internal consistency of neural capacity estimators.

I am extremely grateful for the opportunity, for the constructive discussions and for the help of Professor Poor and Alex Dytso.

[06 October 2026] Update: “Channel Capacity Estimation with Cooperative Neural Architectures: The Role of Objective Functions and f-Divergences” has been accepted at IEEE Transactions on Communications!