Angelos-Nikolaos Kanatas

George Stamatelis, Angelos-Nikolaos Kanatas, George C. Alexandropoulos

Multi-Agent Deep Reinforcement Learning (MADRL) has emerged as a powerful tool for optimizing decentralized decision-making systems in complex settings, such as Dynamic Spectrum Access (DSA). However, deploying deep learning models on resource-constrained edge devices remains challenging due to their high computational cost. To address this challenge, in this paper, we present a novel sparse recurrent MARL framework integrating gradual neural network pruning into the independent actor global critic paradigm. Additionally, we introduce a harmonic annealing sparsity scheduler, which achieves comparable, and in certain cases superior, performance to standard linear and polynomial pruning schedulers at large sparsities. Our experimental investigation demonstrates that the proposed DSA framework can discover superior policies, under diverse training conditions, outperforming conventional DSA, MADRL baselines, and state-of-the-art pruning techniques.

George Stamatelis, Angelos-Nikolaos Kanatas, Ioannis Asprogerakas et al.

Active hypothesis testing is a thoroughly studied problem that finds numerous applications in wireless communications and sensor networks. In this paper, we focus on one centralized and one decentralized problem of active hypothesis testing in the presence of an eavesdropper. For the centralized problem including a single legitimate agent, we present a new framework based on deep NeuroEvolution (NE), whereas, for the decentralized problem, we develop a novel NE-based method for solving collaborative multi-agent tasks, which, interestingly, maintains all computational benefits of our single-agent NE-based scheme. To further reduce the computational complexity of the latter scheme, a novel multi-agent joint NE and pruning framework is also designed. The superiority of the proposed NE-based evasive active hypothesis testing schemes over conventional active hypothesis testing policies, as well as learning-based methods, is validated through extensive numerical investigations in an example use case of anomaly detection over wireless sensor networks. It is demonstrated that the proposed joint optimization and pruning framework achieves nearly identical performance with its unpruned counterpart, while removing a very large percentage of redundant deep neural network weights.