Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · cond-mat/0302536

Phase Diagram for the Constrained Integer Partitioning Problem

C. Borgs·J. T. Chayes·S. Mertens·B. Pittel
View original source ↗Read source document ↗Export metadata ↓

Abstract

We consider the problem of partitioning $n$ integers into two subsets of given cardinalities such that the discrepancy, the absolute value of the difference of their sums, is minimized. The integers are i.i.d. random variables chosen uniformly from the set $\{1,...,M\}$. We study how the typical behavior of the optimal partition depends on $n,M$ and the bias $s$, the difference between the cardinalities of the two subsets in the partition. In particular, we rigorously establish this typical behavior as a function of the two parameters $κ:=n^{-1}\log_2M$ and $b:=|s|/n$ by proving the existence of three distinct ``phases'' in the $κb$-plane, characterized by the value of the discrepancy and the number of optimal solutions: a ``perfect phase'' with exponentially many optimal solutions with discrepancy 0 or 1; a ``hard phase'' with minimal discrepancy of order $Me^{-Θ(n)}$; and a ``sorted phase'' with an unique optimal partition of order $Mn$, obtained by putting the $(s+n)/2$ smallest integers in one subset. Our phase diagram covers all but a relatively small region in the $κb$-plane. We also show that the three phases can be alternatively characterized by the number of basis solutions of the associated linear programming problem, and by the fraction of these basis solutions whose $\pm 1$-valued components form optimal integer partitions of the subproblem with the corresponding weights. We show in particular that this fraction is one in the sorted phase, and exponentially small in both the perfect and hard phases, and strictly exponentially smaller in the hard phase than in the perfect phase. Open problems are discussed, and numerical experiments are presented.

Explore related subjects

cond-mat.dis-nncond-mat.stat-mechcs.CCmath.PR

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

C. Borgs, J. T. Chayes, S. Mertens, B. Pittel. 2003-02-26. Phase Diagram for the Constrained Integer Partitioning Problem. https://arxiv.org/abs/cond-mat/0302536

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Deep-layered machines have a built-in Occam's razor

Many complex systems map vast numbers of microscopic descriptions onto far fewer macroscopic outcomes. Surprisingly, these input-output maps strongly favor simple outputs. In an exactly solvable deep-layered machine, we prove that simplicity bias strengthens with every layer. The probability of an output function $F$ approaches $\Pr(F)\propto 2^{-K(F)}$, where $K$ is the Kolmogorov complexity. Our results show that hierarchical composition alone can generate an intrinsic Occam's razor, suggesting a general mechanism for simplicity bias in learning and evolution.

cond-mat.dis-nn↗
arXiv2026

Neural Networks with Exciton-Polariton Waveguides

We propose and numerically study a scheme for neural networks using exciton-polariton waveguide meshes. These enable arbitrary programmable interlayer couplings suited to a restricted Boltzmann machine architecture where polariton phases act as stochastic network variables. We show that such a network can be trained using unsupervised Hebbian and anti-Hebbian updates. This allows the implementation of an associative memory, where the network recovers learned patterns from imperfect inputs.

cond-mat.dis-nn↗
arXiv2026

Finite-width shift of the edge of chaos in deep neural networks

Deep neural networks train best when they are initialised at the edge of chaos, the critical line in the plane of weight and bias variances on which the distance between two inputs neither grows nor shrinks with depth. The line is known for infinitely wide networks; at finite width it moves, and how it moves has been disputed. We treat the logarithm of the distance between two inputs through a network of width $N$ as a random walk and compute its drift and variance to order $1/N$. The edge of chaos moves to a larger weight variance, by $Δσ_w^2=-a_1/(N\,\partial_{σ_w^2}\lnχ_1)$, about $8/N$ for $\tanh$ networks at a bias variance of $0.05$, confirmed by simulations of the exact finite-width recursion. The shift describes a typical network; the average over many networks is dominated by rare networks in which the distance grows, and its critical point moves the other way, which explains why calculations and simulations have disagreed. Backpropagation follows the same random walk, so the shift applies to gradients unchanged. Along the edge of chaos, the information about the similarity of two inputs is constant until the compression over the depth exceeds about one nat, which fixes where to sit on the line. The results give an initialisation rule, Crinit, that places a network on its finite-width edge of chaos without a search. On MNIST and Fashion-MNIST at width $64$ it is more accurate than the standard rules at their usual settings, by a small margin in shallow networks and a large one in deep networks, where most standard rules fail; it trains in fewer steps, and it matches or beats the standard rules with a tuned gain.

cond-mat.dis-nn↗