Wireless channels in motion-rich urban microcell (UMi) settings are non-stationary; mobility and scatterer dynamics shift the distribution over time, degrading classical and deep estimators. This work proposes conditional prior-based diffusion for channel estimation, which learns a history-conditioned score to denoise noisy channel snapshots. A temporal encoder with cross-time attention compresses a short observation window into a context vector, which captures the channel’s instantaneous coherence and steers the denoiser via feature-wise modulation. During inference, an SNR-matched initialization selects the diffusion step whose marginal aligns with the measured input SNR, and the process follows a shortened, geometrically spaced schedule, preserving the signal-to-noise trajectory with far fewer iterations. Temporal self-conditioning with the previous channel estimate and a training-only smoothness penalty further stabilizes the evolution without biasing the test-time estimator. Evaluations on a 3GPP benchmark show lower NMSE across all SNR regimes than LMMSE, GMM, LSTM, and LDAMP baselines, demonstrating stable performance and strong high-SNR fidelity.

Resource allocation in the multiple-input multiple-output (MIMO) multiple access channel (MAC) is a fundamental problem in multiuser communications, yet it is increasingly treated as non-convex and computationally intractable. This has motivated a large body of heuristic machine learning and successive-approximation methods. Results here show that the MIMO MAC admits canonical convex formulations and present four solvers that together characterize its capacity region. maxRMAC performs weighted sum-rate maximization under per-user energy constraints, minPMAC finds the minimum weighted energy required to support target rates, maxRESMAC performs weighted sum-rate maximization under a total energy constraint, and admMAC tests rate-region feasibility. The solvers exploit the polymatroid structure of the MAC rate region and the separability of the dual Lagrangian across frequency tones, which reduces the problem to parallel per-tone covariance optimizations solved via limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) over Cholesky-like covariance factors. Experiments on spatially correlated MIMO orthogonal frequency-division multiplexing (OFDM) channels show that the proposed solvers match a commercial convex solver in solution quality while running up to two orders of magnitude faster and scaling to regimes where the commercial solver times out. Through broadcast channel (BC) to MAC duality, the same solvers also enable optimal precoder design for the MIMO BC. All solvers are open-sourced and available at https://github.com/muhd-umer/canonical-mac.

To enhance the spectral efficiency and access capacity of communication networks, this work proposes an innovative architecture that integrates rate-splitting multiple access (RSMA) with hybrid beamforming (HBF) for uplink millimeter wave (mmWave) communications. In the RSMA transmission, each user’s message is split into two sub-messages, contributing to the overall transmission rate, and the base station (BS) decodes them using successive interference cancellation (SIC) technique. To maximize the system’s sum-rate, users optimize HBF, while the BS determines the decoding order for all sub-messages. The joint optimization is formulated as a non-convex mixed-integer nonlinear programming (MINLP) problem, subject to power, minimum rate, decoding order, and constant modulus constraints. To meet this challenge efficiently, this work introduces a two-stage optimization framework. The first stage optimizes the analog beamformer, while the second stage utilizes alternating optimization (AO) to optimize the decoding order and digital beamformer. Specifically, for given hybrid beamformers, a low-complexity strategy is derived to obtain the optimal SIC decoding order by ranking the sub-messages according to equivalent channel gains and minimum rate constraints. Furthermore, a fast online-learning-based algorithm for digital beamforming, named Learning-aided differentiable explicit Projection Gradient Descent (LPGD), accelerates the optimization process. Simulation results show that the LPGD-based digital beamforming algorithm converges over 100 times faster than traditional fractional programming (FP)-based methods, while maintaining comparable performance. Moreover, the proposed RSMA scheme achieves up to 26.4% and 34.4% improvement in sum-rate over NOMA and SDMA schemes, respectively.

As wireless networks evolve toward 6G and beyond, deep learning-based channel estimators face a fundamental challenge: their performance degrades when operating conditions deviate from training distributions. This phenomenon, known as distribution shift, manifests through temporal non-stationarity due to user mobility and operational regime changes across varying signal-to-noise ratios (SNR). Traditional sequential training approaches suffer from catastrophic forgetting, where adaptation to new conditions erases previously learned knowledge. This article introduces Agentic Continual Adaptation (ACA), a framework that formulates continual learning as reinforcement-based sequential decision-making to enable continuous adaptation without forgetting. ACA integrates three synergistic mechanisms: an anchor memory that detects forgetting in real-time through continuous performance monitoring on representative samples, a Reinforcement Learning (RL) agent that dynamically selects from seven adaptation strategies balancing plasticity and stability, and a forgetting-aware rollback mechanism that rejects parameter updates exceeding adaptive degradation thresholds. Experimental evaluation across ten SNR regimes demonstrates 4.3% average improvement in channel estimation accuracy with peak gains of 13.6% at challenging low-SNR conditions, while triggering rollback on only 7.5% of parameter updates. This work provides a practical framework for continual adaptation in next-generation wireless networks, addressing an important part of the gap between laboratory performance and real-world deployment.

This letter introduces a Graph-Condensed Quantum-Inspired Placement (GC-QAP) framework for reliability-driven trajectory optimization in Uncrewed Aerial Vehicle (UAV) assisted low-altitude wireless networks. The dense waypoint graph is condensed using probabilistic quantum-annealing to preserve interference-aware centroids while reducing the control state space and maintaining link-quality. The resulting problem is formulated as a priority-aware Markov decision process and solved using epsilon-greedy off-policy Q-learning, considering UAV kinematic and flight corridor constraints. Unlike complex continuous-action reinforcement learning approaches, GC-QAP achieves stable convergence and low outage with substantially and lower computational cost compared to baseline schemes.

Millimeter-wave (mmWave) multi-user MIMO systems are highly susceptible to dynamic blockages, and reconfigurable intelligent surfaces (RIS) have been introduced as a remedy. However, RIS links can themselves be blocked, while existing studies often assume ideal availability. This paper proposes an end-to-end mobility-aware multi-RIS optimization framework that integrates per-RIS blockage detection with closed-form Riemannian updates. The base station transmits short indexed synchronization signals, enabling each user to identify blocked panels via a simple energy test. Based on the detected feasible sets, we jointly optimize the BS precoder and RIS phases using a Stochastic Closed-form Riemannian Phase Alignment (SCRPA) algorithm, which ensures unit-modulus feasibility, monotone convergence, and low complexity. Extensive simulations validate reliable blockage detection and demonstrate significant weighted sum-rate and scalability gains compared to existing baselines.

Large language model reasoning is often treated as a monolithic capability, relying on binary preference supervision that fails to capture partial progress or fine-grained reasoning quality. We introduce Continuous Utility Direct Preference Optimization (CU-DPO), a framework that aligns models to a portfolio of prompt-based cognitive strategies by replacing binary labels with continuous scores that capture fine-grained reasoning quality. We prove that learning with K strategies yields a Theta(K log K) improvement in sample complexity over binary preferences, and that DPO converges to the entropy-regularized utility-maximizing policy. To exploit this signal, we propose a two-stage training pipeline: (i) strategy selection, which optimizes the model to choose the best strategy for a given problem via best-vs-all comparisons, and (ii) execution refinement, which trains the model to correctly execute the selected strategy using margin-stratified pairs. On mathematical reasoning benchmarks, CU-DPO improves strategy selection accuracy from 35-46 percent to 68-78 percent across seven base models, yielding consistent downstream reasoning gains of up to 6.6 points on in-distribution datasets with effective transfer to out-of-distribution tasks.

Accurate and efficient estimation of Channel State Information (CSI) is critical for next-generation wireless systems operating under non-stationary conditions, where user mobility, Doppler spread, and multipath dynamics rapidly alter channel statistics. Conventional pilot aided estimators incur substantial overhead, while deep learning approaches degrade under dynamic pilot patterns and time varying fading. This paper presents a pilot-aided Flash-Attention Transformer framework that unifies model-driven pilot acquisition with data driven CSI reconstruction through patch-wise self-attention and a physics aware composite loss function enforcing phase alignment, correlation consistency, and time frequency smoothness. Under a standardized 3GPP NR configuration, the proposed framework outperforms LMMSE and LSTM baselines by approximately 13 dB in phase invariant normalized mean-square error (NMSE) with markedly lower bit-error rate (BER), while reducing pilot overhead by 16 times. These results demonstrate that attention based architectures enable reliable CSI recovery and enhanced spectral efficiency without compromising link quality, addressing a fundamental bottleneck in adaptive, low-overhead channel estimation for non-stationary 5G and beyond-5G networks.

Modern wireless networks face critical challenges when mobile users traverse heterogeneous network configurations with varying antenna layouts, carrier frequencies, and scattering statistics. Traditional predictors degrade under distribution shift, with NMSE rising by 37.5% during cross-configuration handovers. This work addresses catastrophic forgetting in channel prediction by proposing a continual learning framework based on loss regularization. The approach augments standard training objectives with penalty terms that selectively preserve network parameters essential for previous configurations while enabling adaptation to new environments. Two prominent regularization strategies are investigated: Elastic Weight Consolidation (EWC) and Synaptic Intelligence (SI). Across 3GPP scenarios and multiple architectures, SI lowers the high-SNR NMSE floor by up to 1.8 dB (≈32–34%), while EWC achieves up to 1.4 dB (≈17– 28%). Notably, standard EWC incurs ${\mathcal{O}}(MK)$ complexity (storing M Fisher diagonal entries and corresponding parameter snapshots across K tasks) unless consolidated, whereas SI maintains ${\mathcal{O}}(M)$ memory complexity (storing M model parameters), independent of task sequence length, making it suitable for resource-constrained wireless infrastructure.1

Post-training has split large language model (LLM) alignment into two largely disconnected tracks. Online reinforcement learning (RL) with verifiable rewards drives emergent reasoning on math and code but depends on a programmatic verifier that cannot reach open-ended tasks, while preference optimization handles open-ended generation yet forgoes the continuous exploration that powers online RL. Closing this gap requires a verifier for open-ended quality, but a scalar reward model is the wrong shape for the job. Quality is multi-dimensional, and any scalar score is an incomplete proxy that lets online RL collapse onto whichever axis the score is most sensitive to. We turn instead to the General Preference Model (GPM), which embeds responses into $k$ skew-symmetric subspaces and represents preference as a structured, intransitivity-aware comparison. Building on this, we propose General Preference Reinforcement Learning (GPRL), which carries the $k$-way structure through to the policy update. GPRL computes per-dimension group-relative advantages, normalizes each on its own scale so no axis can dominate, and aggregates them with context-dependent eigenvalues. The same structure powers a closed-loop drift monitor that detects single-axis exploitation and corrects it on the fly by reweighting dimensions and tightening the trust region. Starting from $\texttt{Llama-3-8B-Instruct}$, GPRL reaches a length-controlled win rate of $56.51\%$ on AlpacaEval~2.0 while also outperforming SimPO and SPPO on Arena-Hard, MT-Bench, and WildBench by resisting reward hacking across extended training runs.

Wireless support of virtual reality (VR) has challenges when a network has multiple users, particularly for 3D VR gaming, digital AI avatars, and remote team collaboration. This work addresses these challenges through investigation of the low-rank channels that inevitably occur when there are more active users than there are degrees of spatial freedom, effectively often the number of antennas. The presented approach uses optimal nonlinear transceivers, equivalently generalized decision-feedback or successive cancellation for uplink and superposition or dirty-paper precoders for downlink. Additionally, a powerful optimization approach for the users' energy allocation and decoding order appears to provide large improvements over existing methods, effectively nearing theoretical optima. As the latter optimization methods pose real-time challenges, approximations using deep reinforcement learning (DRL) are used to approximate best performance with much lower (5x at least) complexity. Experimental results show significantly larger sum rates and very large power savings to attain the data rates found necessary to support VR. Experimental results show the proposed algorithm outperforms current industry standards like orthogonal multiple access (OMA), non-orthogonal multiple access (NOMA), as well as the highly researched methods in multi-carrier NOMA (MC-NOMA), enhancing sum data rate by 39%, 28%, and 16%, respectively, at a given power level. For the same data rate, it achieves power savings of 75%,45%, and 40%, making it ideal for VR applications. Additionally, a near-optimal deep reinforcement learning (DRL)-based resource allocation framework for real-time use by being 5x faster and reaching 83% of the global optimum is introduced.

Efficient spectrum allocation has become crucial as the surge in wireless-connected devices demands seamless support for more users and applications, a trend expected to grow with 6G. Innovations in satellite technologies such as SpaceX's Starlink have enabled non-terrestrial networks (NTNs) to work alongside terrestrial networks (TNs) and allocate spectrum based on regional demands. Existing spectrum sharing approaches in TNs use machine learning for interference minimization through power allocation and spectrum sensing, but the unique characteristics of NTNs like varying orbital dynamics and coverage patterns require more sophisticated coordination mechanisms. The proposed work uses a hierarchical deep reinforcement learning (HDRL) approach for efficient spectrum allocation across TN-NTN networks. DRL agents are present at each TN-NTN hierarchy that dynamically learn and allocate spectrum based on regional trends. This framework is 50x faster than the exhaustive search algorithm while achieving 95% of optimum spectral efficiency. Moreover, it is 3.75x faster than multi-agent DRL, which is commonly used for spectrum sharing, and has a 12% higher overall average throughput.

Upcoming Augmented Reality (AR) and Virtual Reality (VR) systems require high data rates (>= 500 Mbps) and low power consumption for seamless experience. With an increasing number of subscribing users, the total number of antennas across all transmitting users far exceeds the number of antennas at the access point (AP). This results in a low rank wireless channel, presenting a bottleneck for uplink communication systems. The current uplink systems that use orthogonal multiple access (OMA) and the proposed non-orthogonal multiple access (NOMA), fail to achieve the required data rates / power consumption under predominantly low rank channel scenarios. This paper introduces an optimal power sub carrier allocation algorithm for multi-carrier NOMA, named minPMAC, and an associated time-sharing algorithm that adaptively changes successive interference cancellation decoding orders to maximize sum data rates in these low rank channels. This Lagrangian based optimization technique, although globally optimum, is prohibitive in terms of runtime, proving inefficient for real-time scenarios. Hence, we propose a novel near-optimal deep reinforcement learning-based energy sum optimization (DRL-minPMAC) which achieves real-time efficiency. Extensive experimental evaluations show that minPMAC achieves 28% and 39% higher data rates than NOMA and OMA baselines. Furthermore, the proposed DRL-minPMAC runs 5 times faster than minPMAC and achieves 83% of the global optimum data rates in real time.

This paper examines the optimal power allocation in a multiple access channel (MAC) and multi-carrier system, subject to a minimum users' rate constraint. The problem is motivated by remote rendering applications with the goal to offload computationally intensive tasks to the mobile edge devices equipped with a non-linear receiver. A shallow-cut Ellipsoid algorithm is proposed to reduce the convergence time of the employed optimization algorithm. The presented idea exploits a key structure present widely in optimization problems involving multi-carrier systems, and demonstrates significant reduction in computational time for finding the solution, with up to 50% fewer iterations. Specifically, the algorithm curtails vertex hopping by an early identification of the time-sharing structure present in a problem instantiation. The accuracy of the final solution is numerically verified.

Emerging wireless AR/VR applications require real-time transmission of correlated high-fidelity speech from multiple resource-constrained devices over unreliable, bandwidth-limited channels. Existing autoencoder-based speech source coding methods fail to address the combination of the following - (1) dynamic bitrate adaptation without retraining the model, (2) leveraging correlations among multiple speech sources, and (3) balancing downstream task loss with realism of reconstructed speech. We propose a neural distributed principal component analysis (NDPCA)-aided distributed source coding algorithm for correlated speech sources transmitting to a central receiver. Our method includes a perception-aware downstream task loss function that balances perceptual realism with task-specific performance. Experiments show significant PSNR improvements under bandwidth constraints over naive autoencoder methods in task-agnostic (19 theoretical upper bound, where all correlated sources are sent to a single encoder, especially in low-bandwidth scenarios. Additionally, we present a rate-distortion-perception trade-off curve, enabling adaptive decisions based on application-specific realism needs.

Intelligent reflecting surfaces (IRSs) improves beyond fifth generation (B5G) systems performance in power- and cost-efficient ways. However, maintaining the performance of multiple IRSs-enabled networks without constraining available resources is challenging. In this paper, we propose a novel IRS-assisted phase cooperative framework to maximize the sum rate of the secondary phase cooperative system (SPC Sys ) located in close proximity of the primary phase cooperative system (PPC Sys ). We exploit transmit beamforming (BF) at base stations (BSs) and phase shift optimization at the IRS with effective phase cooperation between BSs. The maximization problem turns out to be NP-hard, so an alternating optimization is solved for the PPC Sys using an exhaustive search method, i.e., the branch-reduce-and-bound (BRB) algorithm, to obtain the optimal solution for active beamformers, and phase optimization is performed using the semidefinite relaxation (SDR) approach. Further, an active BF is carried out at the SPC Sys transmitter by utilizing optimal phase shifts of the PPC Sys . For the proposed framework, the performance of the BRB algorithm is compared with sub-optimal heuristic BF approaches, including transmit minimum-mean-square-error, zero-forcing BF, and maximum-ratio-transmission. The results support the benefits of deploying IRS in wireless networks to improve sum rate performance of SPC Sys through effective phase cooperation. The proposed framework significantly reduces the hardware cost of the system without constraining the resources of PPC Sys .

Next-generation IoT and cyber-physical systems must deliver fresh updates while keeping queues stable and radios energy-frugal. Existing backlog-centric schedulers minimize delay but ignore data freshness, whereas age-centric policies risk throughput collapse. This work proposes AoI-Queue-Proportional Scheduling (AOI-QPS), an energy-aware extension of classical QPS that allocates service proportional to a weighted sum of queue lengths and Age-of-Information. A Lyapunov-drift analysis proves AOI-QPS is throughput-optimal and bounds average AoI by $\mathcal{O}(1/\alpha )$ while guaranteeing finite delay. Incorporating a IEEE802.15.4-derived energy model further yields an empirical AoI-per-Joule guarantee for queue-stable policies. Extensive simulations under uniform, skewed, and bursty traffic show AOI-QPS cuts network-wide AoI by up to 60% versus QPS, matches Max-AoI’s freshness within 5%, and eliminates the 30% idle-slot energy waste of static schedulers—all without sacrificing queue stability or energy efficiency. AOI-QPS thus addresses the AoI–queue–energy trade-off, offering a practical, analytically grounded knob for balancing freshness, latency, and sustainability in next-generation wireless networks.

This paper proposes a novel joint channel-estimation and source-detection algorithm using successive interference cancellation (SIC)-aided generative score-based diffusion models. Prior work in this area focuses on massive MIMO scenarios, which are typically characterized by full-rank channels, and fail in low-rank channel scenarios. The proposed algorithm outperforms existing methods in joint source-channel estimation, especially in low-rank scenarios where the number of users exceeds the number of antennas at the access point (AP). The proposed score-based iterative diffusion process estimates the gradient of the prior distribution on partial channels, and recursively updates the estimated channel parts as well as the source. Extensive simulation results show that the proposed method outperforms the baseline methods in terms of normalized mean squared error (NMSE) and symbol error rate (SER) in both full-rank and low-rank channel scenarios, while having a more dominant effect in the latter, at various signal-to-noise ratios (SNR).

Future wireless networks aims to deliver high data rates and lower power consumption while ensuring seamless connectivity, necessitating robust wireless network optimization. To achieve this, wireless network optimization is necessary. Large language models (LLMs) have been deployed for generalized optimization scenarios. To take advantage of generative AI (GAI) models, retrieval augmented generation (RAG) is proposed for multi-sensor wireless environment perception. Utilizing domainspecific prompt engineering, we apply retrieval-augmented generation (RAG) to efficiently harness multimodal data inputs from sensors in a wireless environment. Wireless environment perception is necessary for global LLM optimization tasks. Key pre-processing pipelines including image-to-text conversion, object detection, and distance calculations for multimodal RAG input from multi-sensor data from different devices are proposed in this paper to obtain a unified vector database crucial for optimizing large language models (LLMs) in global wireless tasks. Our evaluation, conducted with OpenAI's GPT and Google's Gemini models, demonstrates an 8%, 8%, 10%, 7%, and 12% improvement in relevancy, faithfulness, completeness, similarity, and accuracy, respectively, compared to conventional LLM-based designs. Furthermore, our RAG-based LLM framework with vectorized databases are computationally efficient providing real time convergence under latency constraints.

We investigate the capacity region of multi-user interference channels (IC), where each user encodes multiple sub-user components. By unifying chain-rule decomposition with the Entropy Power Inequality (EPI), we reason that single-user Gaussian codebooks suffice to achieve optimal performance, thus obviating any need for intricate auxiliary variables or joint typicality arguments. Our partial-MAC formulation enumerates sub-user decoding orders while only imposing constraints for sub-users actually decoded. This significantly reduces complexity relative to enumerating all subsets or bruteforcing over all successive interference cancellation (SIC) decoding order combinations at all receivers. This leads to a finite but comprehensive construction of all achievable rate tuples under sum-power constraints, while guaranteeing that each receiver fully recovers its intended sub-user signals. Consequently, known single-user Gaussian capacity results generalize naturally to multi-user scenarios, revealing a cohesive framework for analyzing multi-user IC. Our results thus offer a streamlined, tractable pathway for designing next-generation cell-free wireless networks that rely on IC mechanisms, efficiently exploiting interference structure while minimizing overhead. Overall, this provides a unifying perspective.