156 papers
This paper proposes Belief-Adaptive MAP and Soft BA-MAP detectors that explicitly account for state-dependent noise variance in molecular communication channels, demonstrating significant throughput improvements over conventional methods through simulations and information-theoretic analysis.
This paper analyzes the secrecy performance of continuous-aperture array (CAPA) systems over Rayleigh fading channels, deriving theoretical expressions for secrecy rate and outage probability across single, multiple independent, and multiple collaborative eavesdropper scenarios, and demonstrating that CAPA systems outperform traditional spatially-discrete arrays by achieving higher secrecy rates, lower outage probabilities, and a diversity order equal to the spatial degrees of freedom.
This paper proposes a distributed non-cooperative game-theoretic framework for jointly optimizing linear MIMO transceivers to resolve semantic mismatches and mitigate multi-user interference in goal-oriented communication systems, deriving a closed-form Nash Equilibrium that balances information compression, interference management, and task performance.
This paper proposes a physically consistent signal model and an alternating optimization framework for downlink sum-rate maximization in STAR beyond-diagonal RISs with per-element amplification, achieving substantial performance gains over conventional passive systems while ensuring hardware feasibility and passivity.
This paper proposes a joint optimization algorithm for transmit precoding and discrete phase-shift reflection in an extremely large-scale RIS-aided near-field secure communication system with artificial jamming, effectively maximizing secrecy rates even when eavesdroppers are located close to the RIS and in the same direction as legitimate users.
This paper introduces a unified framework for learning unknown objects from arbitrary linear samples using general nonlinear model classes, establishing near-optimal generalization bounds based on the model's variation and complexity while recovering and extending existing results in areas like compressed sensing and matrix sketching.
This paper introduces Speculative Decoding Scaling Laws (SDSL), a theoretical framework that analytically links pre-trained LLM hyperparameters to inference throughput, enabling the prediction of optimal system configurations without costly experimental training.
This paper proves that super two-extendible states, a broad class including erased and full-rank states, cannot be used for heralded exact one-way secret-key distillation, thereby revealing a significant gap between exact and approximate distillable key rates.
This paper extends the Ziv-Merhav theorem on universal estimation of specific cross-entropy from multi-level Markov measures to a broader class of decoupled measures, including regular g-measures and equilibrium measures from statistical mechanics.
This paper proposes an energy efficiency maximization framework for continuous aperture array (CAPA)-based multi-group multicast communications by developing an optimal block coordinate descent algorithm and a low-complexity zero-forcing approach, demonstrating CAPA's superior performance over discrete arrays while revealing nuanced impacts of aperture size and user distribution.
This paper proposes a discrete mathematical framework using coarse-grained partitions and categorical unification to quantify information loss in explainable AI, demonstrating that zero loss is an exceptional case and providing a method to optimize the trade-off between interpretability and informational fidelity.
This paper explores the application of Generative Diffusion Models (GDMs) in massive MIMO communications by reviewing their theoretical foundations, analyzing recent advancements including a case study on near-field channel estimation, and outlining future research directions and challenges for enhancing channel state information acquisition in 5G and 6G networks.
This paper establishes the theoretical framework for solving axial symmetric Navier-Stokes equations in a cylindrical topology by constructing a complete basis of harmonic 1-forms comprising Beltrami, anti-Beltrami, and closed components, thereby reducing the problem to a hierarchy of quadratic relations suitable for future optimization via Physics-Informed Neural Networks.
This paper proposes and analytically evaluates a spatially-aware secondary license sharing framework for mmWave networks that leverages the directional nature of signals and blockage conditions to optimize spectrum sharing, demonstrating that these physical characteristics can significantly enhance transmission opportunities for secondary users while maintaining primary network performance.
This paper establishes that the optimal rebalancing trajectory for Dynamic Weight Automated Market Makers corresponds to a geodesic under the Fisher-Rao metric (SLERP in Hellinger coordinates), thereby proving that the recursive AM-GM bisection heuristic used in prior work lies exactly on this loss-minimizing path.
This paper proposes a cost-effective DNA storage system for progressive images that leverages Nanopore adaptive sampling to enable PCR-free random access, allowing users to retrieve specific resolution layers by selectively sequencing only the necessary oligos rather than the entire pool.
This paper provides a precise asymptotic analysis of multi-task learning in misspecified perceptron models, demonstrating that combining related tasks is equivalent to adding implicit regularization that improves generalization and mitigates the double descent phenomenon.
This paper presents a formal control synthesis framework for continuous-state stochastic systems that minimizes a linear combination of system entropy (measured by KL divergence to uniform) and cumulative cost by deriving novel bounds on the entropy difference between continuous distributions and their finite-state abstractions, thereby enabling entropy-aware controllers with rigorous performance guarantees.
This paper investigates how mismatched log-likelihood ratios (LLRs) impact belief propagation decoding of overcomplete quantum LDPC codes, revealing that while performance is sensitive to initialization in low noise, the optimal FER is robust across a wide range of LLR values, suggesting that LLR mismatch can be effectively utilized as a regularization control parameter rather than a quantity requiring precise channel matching.
This paper introduces WaterSIC, a novel linear layer quantization algorithm that achieves information-theoretically near-optimal performance by allocating different quantization rates to weight columns via a waterfilling strategy, thereby significantly outperforming existing methods like GPTQ and establishing new state-of-the-art results for LLMs across 1 to 4-bit quantization rates.