Hephaestus: Mixture Generative Modeling with Energy Guidance for Large-scale QoS Degradation

2025-10-19

AI Summary

This paper addresses the Quality-of-Service Degradation (QoSD) problem—where an adversary degrades network performance by perturbing edge weights—posing a core challenge in generating effective attack strategies for large-scale networks under nonlinear edge-weight functions. Existing approaches are limited by combinatorial optimization complexity, restrictive linearity assumptions, or scalability bottlenecks. To overcome these, we propose the first generative, energy-guided self-reinforcing framework for QoSD: it couples a hybrid conditional variational autoencoder with an energy-based model, integrates a predictive path-pressure algorithm, and employs differentiable reward-driven reinforcement learning—enabling end-to-end, differentiable, and scalable QoSD solving. Evaluated on both synthetic and real-world networks, our method significantly outperforms classical combinatorial optimization and state-of-the-art ML baselines, especially under nonlinear cost functions, demonstrating strong generalization, computational efficiency, and practical deployability.

Technology Category

• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Adversarial Learning & Robustness

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

We study the Quality of Service Degradation (QoSD) problem, in which an adversary perturbs edge weights to degrade network performance. This setting arises in both network infrastructures and distributed ML systems, where communication quality, not just connectivity, determines functionality. While classical methods rely on combinatorial optimization, and recent ML approaches address only restricted linear variants with small-size networks, no prior model directly tackles the QoSD problem under nonlinear edge-weight functions. This work proposes PIMMA, a self-reinforcing generative framework that synthesizes feasible solutions in latent space, to fill this gap. Our method includes three phases: (1) Forge: a Predictive Path-Stressing (PPS) algorithm that uses graph learning and approximation to produce feasible solutions with performance guarantee, (2) Morph: a new theoretically grounded training paradigm for Mixture of Conditional VAEs guided by an energy-based model to capture solution feature distributions, and (3) Refine: a reinforcement learning agent that explores this space to generate progressively near-optimal solutions using our designed differentiable reward function. Experiments on both synthetic and real-world networks show that our approach consistently outperforms classical and ML baselines, particularly in scenarios with nonlinear cost functions where traditional methods fail to generalize.

Problem

• Addressing Quality of Service Degradation via adversarial edge weight perturbations
• Solving nonlinear edge-weight QoSD problems in large-scale networks
• Generating feasible near-optimal solutions for network performance degradation

Innovation

• Predictive Path-Stressing algorithm generates feasible solutions
• Mixture of Conditional VAEs guided by energy-based model
• Reinforcement learning agent refines solutions using differentiable rewards