Semantic and task-oriented communications have emerged as significant paradigm shifts for next-generation communication networks, which extracts and transmits task-relevant information rather than raw data for downstream tasks. However, most existing work focused on bit-level loss functions, such as mean square error (MSE) and cross-entropy (CE), rather than directly optimizing at the semantic level. These approaches often lack interpretability of semantic representation and result in higher system complexity and less efficient transmission for various tasks. To this end, we develop a novel task-oriented semantic communication system for multitask scenarios and further develop a semantic-level framework. This framework can extract structured semantic representation by compressing raw data with different labels into mutually orthogonal subspaces. Simulation results demonstrate that the proposed framework not only extracts structured semantic representation, but also outperforms existing benchmarks in terms of data recovery and AI inference performance.

Massive access is a critical design challenge of Internet of Things (IoT) networks. In this paper, we consider the grant-free uplink transmission of an IoT network with a multiple-antenna base station (BS) and a large number of single-antenna IoT devices. Taking into account the sporadic nature of IoT devices, we formulate the joint activity detection and channel estimation (JADCE) problem as a group-sparse matrix estimation problem. This problem can be solved by applying the existing compressed sensing techniques, which however either suffer from high computational complexities or lack of algorithm robustness. To this end, we propose a novel algorithm unrolling framework based on the deep neural network to simultaneously achieve low computational complexity and high robustness for solving the JADCE problem. Specifically, we map the original iterative shrinkage thresholding algorithm (ISTA) into an unrolled recurrent neural network (RNN), thereby improving the convergence rate and computational efficiency through end-to-end training. Moreover, the proposed algorithm unrolling approach inherits the structure and domain knowledge of the ISTA, thereby maintaining the algorithm robustness, which can handle non-Gaussian preamble sequence matrix in massive access. With rigorous theoretical analysis, we further simplify the unrolled network structure by reducing the redundant training parameters. Furthermore, we prove that the simplified unrolled deep neural network structures enjoy a linear convergence rate. Extensive simulations based on various preamble signatures show that the proposed unrolled networks outperform the existing methods in terms of the convergence rate, robustness and estimation accuracy.

Intelligent Reflecting Surface (IRS) is a promising technology that is able to manipulate the wireless propagation channels via smartly adjusting the signal reflection. With continuous phase shifts, IRS has been shown to be effective in enlarging the achievable rate region. In this paper, we investigate the achievable rate region of a IRS-aided multi-user interference channel, where the phase shifts at the IRS can only take a finite number of discrete values. We formulate a multi-objective optimization problem (MOOP) to characterize the achievable rate region. The commonly adopted approaches such as the rate profile method fail to solve MOOP with optimization variables. Although the exhaustive search method can obtain the Pareto-optimal solutions, it suffers from high computational complexity. To this end, we propose a computationally efficient active learning algorithm via Gaussian process (GP). By modeling the objectives of MOOP as a draw from a GP distribution with only a few randomly computed rate-tuples, the active learning algorithm can quickly dominate the non-optimal points and find Pareto-optimal points without calculating rate-tuples. Numerical simulations demonstrate that the achievable rate region of IRS-aided interference channel is much larger than that without IRS and the proposed active learning framework obtains near-optimal Pareto solutions with a much lower computational complexity than the traditional exhaustive search algorithm.

In this paper, we consider decentralized federated learning (FL) over wireless networks, where over-the-air computation (AirComp) is adopted to facilitate the local model consensus in a device-to-device (D2D) communication manner. However, the AirComp-based consensus phase brings the additive noise in each algorithm iterate and the consensus needs to be robust to wireless network topology changes, which introduce a coupled and novel challenge of establishing the convergence for wireless decentralized FL algorithm. To facilitate consensus phase, we propose an AirComp-based DSGD with gradient tracking and variance reduction (DSGT-VR) algorithm, where both precoding and decoding strategies are developed for D2D communication. Furthermore, we prove that the proposed algorithm converges linearly and establish the optimality gap for strongly convex and smooth loss functions, taking into account the channel fading and noise. The theoretical result shows that the additional error bound in the optimality gap depends on the number of devices. Extensive simulations verify the theoretical results and show that the proposed algorithm outperforms other benchmark decentralized FL algorithms over wireless networks.