Abstract
We consider the symmetric binary perceptron model, a simple model of neuralnetworks that has gathered significant attention in the statistical physics,information theory and probability theory communities, with recent connectionsmade to the performance of learning algorithms in Baldassi et al. '15. We establish that the partition function of this model, normalized by itsexpected value, converges to a lognormal distribution. As a consequence, thisallows us to establish several conjectures for this model: (i) it proves thecontiguity conjecture of Aubin et al. '19 between the planted and unplantedmodels in the satisfiable regime; (ii) it establishes the sharp thresholdconjecture; (iii) it proves the frozen 1-RSB conjecture in the symmetric case,conjectured first by Krauth-M\'ezard '89 in the asymmetric case. In a recent concurrent work of Perkins-Xu [PX21], the last two conjectureswere also established by proving that the partition function concentrates on anexponential scale. This left open the contiguity conjecture and the lognormallimit characterization, which are established here. In particular, our prooftechnique relies on a dense counter-part of the small graph conditioningmethod, which was developed for sparse models in the celebrated work ofRobinson and Wormald.