Enhancement of Over-the-Air Federated Learning by Using AI-based Fluid Antenna System

This letter investigates an over-the-air federated learning (OTA-FL) system that employs fluid antennas (FAs) at the access point (AP) to enhance learning performance by leveraging the additional degrees of freedom provided by antenna mobility. First, we analyze the convergence of the OTA-FL system and derive the optimality gap to illustrate the influence of FAs on learning performance. Then, we formulate a nonconvex optimization problem to minimize the optimality gap by jointly optimizing the positions of the FAs and the beamforming vector. To address the dynamic environment, we cast this optimization problem as a Markov decision process (MDP) and propose the recurrent deep deterministic policy gradient (RDPG) algorithm. Finally, extensive simulations show that the FA-assisted OTA-FL system outperforms systems with fixed-position antennas and that the RDPG algorithm surpasses the existing methods.

A2P-MANN: Adaptive Attention Inference Hops Pruned Memory-Augmented Neural Networks

In this work, to limit the number of required attention inference hops in memory-augmented neural networks, we propose an online adaptive approach called A2P-MANN. By exploiting a small neural network classifier, an adequate number of attention inference hops for the input query is determined. The technique results in elimination of a large number of unnecessary computations in extracting the correct answer. In addition, to further lower computations in A2P-MANN, we suggest pruning weights of the final FC (fully-connected) layers. To this end, two pruning approaches, one with negligible accuracy loss and the other with controllable loss on the final accuracy, are developed. The efficacy of the technique is assessed by using the twenty question-answering (QA) tasks of bAbI dataset. The analytical assessment reveals, on average, more than 42% fewer computations compared to the baseline MANN at the cost of less than 1% accuracy loss. In addition, when used along with the previously published zero-skipping technique, a computation count reduction of up to 68% is achieved. Finally, when the proposed approach (without zero-skipping) is implemented on the CPU and GPU platforms, up to 43% runtime reduction is achieved.