Message-passing graph neural networks (MP-GNNs) are widely used for learning on relational data. However, their performance drops on tasks requiring long-range interactions due to over-squashing, where exponential information compression overwhelms fixed-width embeddings. While existing analyses often attribute this to geometric bottlenecks under linear diffusion assumptions, thresholded nonlinearities in GNNs motivate a load-release view akin to Abelian sandpiles. Using the discrete sandpile model as a structural proxy, we show that graph bottlenecks force large stabilization cost, effectively creating zones of high transport congestion. We characterize stabilization-invariant equivalence classes induced by the reduced Laplacian and derive cut-based lower bounds linking bottlenecks to unavoidable stabilization effort. The resulting theory is discrete, whereas our implementation is a continuous vector-valued surrogate. The theory identifies the relevant design factors, namely capacity and cut size. Guided by these insights, we propose a differentiable Sandpile Stabilization Layer (SSL) and congestion-aware objectives designed to redistribute excess load and manage stabilization costs. Experiments on long-range benchmarks, together with congestion and collision diagnostics, show that targeting sandpile-identified bottlenecks mitigates representation collapse and improves over standard baselines. Project Page: https://sandpile-gnn.github.io/
Yang Shi, Lixian Chen, Jingchao Wang et al.· Proceedings of the 32nd ACM...· 0 citations
The Traveling Salesman Problem (TSP) is a classic NP-hard combinatorial optimization problem where traditional intelligent algorithms often suffer from slow convergence and premature convergence to local optima. This paper proposes a novel hybrid optimization framework (HGA-PSO-ACO) that integrates the global exploration capability of the Genetic Algorithm (GA), the rapid convergence characteristics of Particle Swarm Optimization (PSO), and the positive feedback mechanism of Ant Colony Optimization (ACO) through a three-stage adaptive coordination model. The key design challenge is to determine when each component should dominate the search and how information should be transferred across stages without causing premature homogenization of the population. To address this issue, the algorithm employs fitness variance (FVD) for real-time population state monitoring and implements dynamic algorithm switching strategies. Experimental validation on a 40-city TSP benchmark demonstrates that the hybrid algorithm achieves a 40% faster convergence speed, a 2.1% lower relative error, and a 58.3% reduction in run-to-run standard deviation compared to individual algorithms, indicating improved solution stability together with favorable parameter sensitivity control.
Xuhang Chen, Zheng-Rui Zhang, Ze-Hong Chen et al.· 2026 3rd World Conference on...· 0 citations
This approach combines local editing and background preservation without external or user-provided spatial masks and without model fine-tuning, and achieves the strongest reported preservation metrics, including 27.44~dB PSNR and 0.055 LPIPS.
Yang Shi, Liangsi Lu, Minzhe Guo et al.· 0 citations