Paper “Model-Free Learning of Two-Stage Beamformers for Passive IRS-Aided Network Design” has been accepted to IEEE Transactions on Signal Processing.