arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2605.29005cs.LGcs.AI
收藏

LoRe: 基于每步交互预算的自适应交互评估路由用于迭代图求解器

LoRe: Adaptive Interaction-Evaluation Routing with Per-Step Interaction Budgets for Iterative Graph Solvers

Beijing Key Laboratory of Fault-Tolerant Quantum Computing, Beijing Academy of Quantum Information Sciences, Beijing, China(北京容错量子计算重点实验室,量子信息科学北京市院,北京,中国);Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, CAS, Beijing, China(北京凝聚态物理国家实验室,物理研究所,中国科学院,北京,中国);Beijing Key Laboratory of Advanced Quantum Technology, Beijing, China(北京先进量子技术重点实验室,北京,中国);Hefei National Laboratory, Hefei, China(合肥国家实验室,合肥,中国)

Jintao Li, Yong-Yi Wang, Zheng-An Wang, Heng Fan

AI总结:

提出LoRe方法,通过动态路由计算到高冲突或高不确定性交互,实现每步固定比例交互评估,在不牺牲解质量的前提下显著提升迭代图求解器的可扩展性和速度。

摘要:

基于扩散的组合优化神经求解器反复重新评估密集的边/因子交互,导致推理时间昂贵且在大规模下常受内存限制。受多体物理计算方法的启发,我们引入LoRe,一种无需训练、推理时即插即用的包装器,强制执行每步交互评估预算:在每次迭代中,它通过动态路由计算到高冲突或高不确定性交互,仅评估固定比例的交互,而不是使用固定的稀疏化(例如静态kNN图或静态掩码)。在完全包含的端到端挂钟时间核算下,LoRe显著提高了最大独立集(MIS)问题的可扩展性,将可行推理扩展到基线内存溢出限制的3倍以上,实现了约8倍的加速和约12倍的峰值内存减少,同时在此范围内保持解质量。在大规模旅行商问题(TSP)上展示了跨任务通用性,并对拓扑变化具有零样本鲁棒性,LoRe在n=1000时实现了约15倍的加速,内存减少44倍,且巡回质量具有竞争力。

英文摘要:

Diffusion-based neural solvers for combinatorial optimization repeatedly re-evaluate dense edge/factor interactions, making inference expensive in wall-clock time and often memory-bound at scale. Inspired by the computational methodologies of many-body physics, we introduce LoRe, a training-free, inference-time drop-in wrapper that enforces per-step interaction-evaluation budgeting: at each iteration, it evaluates only a fixed fraction of interactions by dynamically routing computation to high-conflict or high-uncertainty interactions, instead of using a fixed sparsification (e.g., static kNN graphs or static masks). Under fully inclusive end-to-end wall-clock accounting, LoRe substantially improves scalability on the Maximum Independent Set (MIS) problem, extending feasible inference more than $3\times$ beyond the baseline's out-of-memory limit, delivering a $\sim 8\times$ speedup and a $\sim 12\times$ peak-memory reduction, with solution quality preserved in this regime. Demonstrating cross-task generality on the large-scale Traveling Salesperson Problem (TSP) and zero-shot robustness to topology shifts, LoRe achieves a $\sim 15\times$ speedup at $n=1000$ with a $44\times$ memory reduction and competitive tour quality.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑