New WGAN framework uses RL and game theory for resource-constrained vehicles

Researchers have developed an adaptive multi-discriminator Wasserstein GAN (MD-WGAN) framework designed for resource-constrained Internet of Vehicles (IoV) environments. This framework integrates reinforcement learning with game theory to manage machine learning workloads as a network service, addressing challenges like dynamic topologies and competing optimization objectives. The system uses roadside units with generators and Deep Q-Network agents to select discriminators and manage distributed training, while a game-theoretic coordination allocates training epochs. Evaluations on NGSIM trajectory data show the MD-WGAN achieves prediction accuracy comparable to state-of-the-art GANs, with significantly improved resource efficiency, including reduced CPU utilization and memory usage. AI

IMPACT This research could enable more efficient deployment of advanced AI models in mobile and resource-limited environments like connected vehicles.

RANK_REASON Academic paper detailing a novel framework and its evaluation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Deep Q-Network
• Farhoud Jafari Kaleibar
• game theory
• generative adversarial network
• Internet of Vehicles
• MD-WGAN
• NGSIM
• reinforcement learning
• WGAN

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →