A Generative Neural Annealer for Black-Box Combinatorial Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Yuan-Hang, Di Ventra, Massimiliano
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918115622780928
author Zhang, Yuan-Hang
Di Ventra, Massimiliano
author_facet Zhang, Yuan-Hang
Di Ventra, Massimiliano
contents We propose a generative, end-to-end solver for black-box combinatorial optimization that emphasizes both sample efficiency and solution quality on NP problems. Drawing inspiration from annealing-based algorithms, we treat the black-box objective as an energy function and train a neural network to model the associated Boltzmann distribution. By conditioning on temperature, the network captures a continuum of distributions--from near-uniform at high temperatures to sharply peaked around global optima at low temperatures--thereby learning the structure of the energy landscape and facilitating global optimization. When queries are expensive, the temperature-dependent distributions naturally enable data augmentation and improve sample efficiency. When queries are cheap but the problem remains hard, the model learns implicit variable interactions, effectively "opening" the black box. We validate our approach on challenging combinatorial tasks under both limited and unlimited query budgets, showing competitive performance against state-of-the-art black-box optimizers.
format Preprint
id arxiv_https___arxiv_org_abs_2505_09742
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Generative Neural Annealer for Black-Box Combinatorial Optimization
Zhang, Yuan-Hang
Di Ventra, Massimiliano
Machine Learning
Disordered Systems and Neural Networks
Statistical Mechanics
Artificial Intelligence
Neural and Evolutionary Computing
We propose a generative, end-to-end solver for black-box combinatorial optimization that emphasizes both sample efficiency and solution quality on NP problems. Drawing inspiration from annealing-based algorithms, we treat the black-box objective as an energy function and train a neural network to model the associated Boltzmann distribution. By conditioning on temperature, the network captures a continuum of distributions--from near-uniform at high temperatures to sharply peaked around global optima at low temperatures--thereby learning the structure of the energy landscape and facilitating global optimization. When queries are expensive, the temperature-dependent distributions naturally enable data augmentation and improve sample efficiency. When queries are cheap but the problem remains hard, the model learns implicit variable interactions, effectively "opening" the black box. We validate our approach on challenging combinatorial tasks under both limited and unlimited query budgets, showing competitive performance against state-of-the-art black-box optimizers.
title A Generative Neural Annealer for Black-Box Combinatorial Optimization
topic Machine Learning
Disordered Systems and Neural Networks
Statistical Mechanics
Artificial Intelligence
Neural and Evolutionary Computing
url https://arxiv.org/abs/2505.09742