Breaking Local-Minimum Traps in Spiking Neural Network-Based Solvers for CSPs via Parallel Tempering

Recep Bugra Uludag; Ahmet Efe; Ismail Akturk; Ulya R. Karpuzcu

Home/Archive/arXiv

arXiv · 2607.08897

Breaking Local-Minimum Traps in Spiking Neural Network-Based Solvers for CSPs via Parallel Tempering

Recep Bugra Uludag·Ahmet Efe·Ismail Akturk·Ulya R. Karpuzcu

View original source ↗Read source document ↗Export metadata ↓

Abstract

Spiking neural networks (SNNs) with stochastic neurons can solve constraint satisfaction problems (CSPs) by encoding constraints via connectivity and performing probabilistic search via spike dynamics. However, fixed-temperature stochastic dynamics often get trapped in local minima - near-satisfying configurations - a vulnerability that escalates with problem difficulty. To overcome this, we integrate parallel tempering (PT) into the neural sampling solver, running multiple parallel replicas at varying inverse temperatures. Replicas periodically exchange temperatures rather than network states, managing the trade-off between exploration and concentration around low-energy configurations while preserving asynchronous, spike-based computation. We evaluate this architecture against a parallel baseline of four independent, fixed-temperature solvers using equal computational resources across 1000 instances from the SATLIB uf20-91 benchmark. Parallel tempering improves success probability on 332 instances while worsening only 5. Crucially, these gains are concentrated on hard instances where independent solvers fail. Violation trajectory analysis confirms the underlying mechanism: temperature exchanges allow replicas to traverse energy barriers unreachable by fixed-temperature dynamics, successfully escaping the narrow basins that constrain the baseline. To our knowledge, this represents the first integration of parallel tempering into an SNN-based CSP solver.

Explore related subjects

cs.ET

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Recep Bugra Uludag, Ahmet Efe, Ismail Akturk, Ulya R. Karpuzcu. 2026-07-09. Breaking Local-Minimum Traps in Spiking Neural Network-Based Solvers for CSPs via Parallel Tempering. https://arxiv.org/abs/2607.08897

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

KEEP EXPLORING

Related papers

arXiv2026

Beyond HBM-on-GPU: Thermal Design Envelope for 3D Volumetric DRAM-on-GPU Integration

The scaling of GPUs for AI and HPC workloads is increasingly constrained by the capacity, bandwidth, and thermal limits of both 2.5D HBM-GPU and direct-stacked 3D HBM-on-GPU integration. This work establishes the thermal design envelope for 3D volumetric DRAM-on-GPU integration, in which vertically oriented DRAM dies and interleaved cooling cavities reshape heat flow and memory interfacing above the GPU. Using a package-level thermal model anchored to a consistent HBM-on-GPU baseline and driven by a realistic reticle-scale non-uniform GPU power map, we quantify the key parameters governing thermal feasibility. Stack height is the dominant limiter of peak temperature, while cooling-cavity conductivity shifts the feasible region, and mold insertion and stack orientation further modulate thermal behavior. A distributed memory-controller and network-on-chip tier introduces only a moderate thermal penalty. Although die-level parallelism increases bandwidth, the reduction in simulated training time saturates once execution becomes compute-bound. These results define a bounded co-design space across bandwidth, capacity, and thermal constraints for 3D volumetric DRAM-on-GPU integration.

cs.ET↗

arXiv2026

Droop-Aware Foundation Model Power Flow

This paper develops a droop-aware extension of the GridFM power systems foundation model, embedding droop gains and frequency/voltage deadband parameters as per-bus node features to enable control-aware AC power-flow analysis. Existing power-flow datasets encode only static electrical features, conflating operating points from qualitatively different control regimes; this work resolves that gap by exposing droop and deadband parameters as structured node features, with deadband discontinuities handled through a smooth tanh approximation that preserves solver differentiability. A transformer-based graph neural network is pre-trained on masked reconstruction and fine-tuned on the resulting control-aware datasets. The framework is validated against PSCAD electromagnetic-transient simulations on a two-bus system (0.11% maximum steady-state error) and cross-validated against an independent PyPower droop solver on the IEEE 24-bus RTS. On the 24-bus system the surrogate attains R2 = 0.9996 for active generation and 0.0015 p.u. voltage-magnitude RMSE; scalability is confirmed on the IEEE 300-bus system (0.0036 p.u. RMSE, R2 = 0.9841 for voltage magnitude across 299,700 predictions). A three-mode control study further shows that the deadband widens the control-error distribution while leaving total droop compensation unchanged, establishing deadband width as an actionable node-level design feature.

cs.ET↗

arXiv2026

GNN-based Path-aware multi-view Circuit Learning for Technology Mapping

Traditional technology mapping suffers from systemic inaccuracies in delay estimation due to its reliance on abstract, technology-agnostic delay models that fail to capture the nuanced timing behavior behavior of real post-mapping circuits. To address this fundamental limitation, we introduce GPA(graph neural network (GNN)-based Path-Aware multi-view circuit learning), a novel GNN framework that learns precise, data-driven delay predictions by synergistically fusing three complementary views of circuit structure: And-Inverter Graphs (AIGs)-based functional encoding, post-mapping technology emphasizes critical timing paths. Trained exclusively on real cell delays extracted from critical paths of industrial-grade post-mapping netlists, GPA learns to classify cut delays with unprecedented accuracy, directly informing smarter mapping decisions. Evaluated on the 19 EPFL combinational benchmarks, GPA achieves 19.9%, 2.1% and 4.1% average delay reduction over the conventional heuristics methods (techmap, MCH) and the prior state-of-the-art ML-based approach SLAP, respectively-without compromising area efficiency.

cs.ET↗