4 papers · 1 filter
Graph Neural Networks are Heuristics
Yimeng Min, Carla P. Gomes
Graph neural networks are usually treated as auxiliaries for combinatorial optimization: they imitate algorithms, guide search, or supply scores to classical procedures. We show th…
Divergence-Suppressing Couplings for Rectified Flow
Yimeng Min, Carla P. Gomes
The promise of Rectified Flow rests on producing self-generated couplings whose trajectories are straight, or nearly so. In practice, trajectories generated by the base flow model…
Unsupervised Learning for Quadratic Assignment
Yimeng Min, Carla P. Gomes
We introduce PLUME search, a data-driven framework that enhances search efficiency in combinatorial optimization through unsupervised learning. Unlike supervised or reinforcement l…
On Size and Hardness Generalization in Unsupervised Learning for the Travelling Salesman Problem
Yimeng Min, Carla P. Gomes
We study the generalization capability of Unsupervised Learning in solving the Travelling Salesman Problem (TSP). We use a Graph Neural Network (GNN) trained with a surrogate loss…