Fig. 5From: Joint high-dimensional soft bit estimation and quantization using deep learningImpact of latent feature dimension on different instances of EQ-Net without numerical quantization in an i.i.d. Rayleigh fading, \(2 \times 2\), 64-QAM scenario using an (324, 648) LDPC code. A feature dimension of six corresponds to the upper bound in Theorem 2, which is proven explicitly for diagonalized channels, but is shown here to be practically achievable for arbitrary channelsBack to article page