With the emergence of three-dimensional Gaussian Splatting (3DGS) in extended reality (XR), the resulting surge in data volume poses a critical challenge to achieving high-quality rendering under limited communication and computation resources. This paper proposes an innovative distributed rendering framework, establishing a unified paradigm for resource allocation in 3DGS-based XR systems. To overcome the challenges of distributed rendering while simultaneously satisfying user’s stringent requirements of users, we formulate an energy-minimization problem for a near-field XR system. Specifically, by jointly optimizing model selection, downlink transmission precoding, and rendering tasks offloading, we aim at minimizing the total energy consumption of system subject to the constraints of end-to-end latency and rendering quality. To address this mixed-integer and nonconvex design problem, we develop an efficient algorithm capitalizing on the block coordinate descent framework for obtaining an effective solution. Simulation results demonstrate that our proposed framework substantially reduces the energy consumption of 3DGS-driven XR systems, providing a paradigm for resource allocation of distributed 3D reconstruction.

Utilizing complex-valued neural networks (CVNNs) in wireless communication tasks has received growing attention for their ability to provide natural and effective representation of complex-valued signals and data. However, existing studies typically employ complex-valued versions of simple neural network architectures. Not only they merely scratch the surface of the extensive range of modern deep learning techniques, theoretical understanding of the superior performance of CVNNs is missing. To this end, this paper aims to fill both the theoretical and practice gap of employing CVNNs in wireless communications. In particular, we provide a comprehensive description on the various operations in CVNNs and theoretically prove that when the output dimension is large, the CVNN requires fewer layers than its real-valued counterpart to achieve a given approximation error of a continuous complex-valued function. Furthermore, to advance CVNNs in the field of wireless communications, this paper focuses on the transformer model, which represents a more sophisticated deep learning architecture and has been shown to have excellent performance in wireless communications but only in its real-valued form. In this aspect, we propose a fundamental paradigm of complex-valued transformers for wireless communications, including the complex-valued embedding module, encoding module, decoding module, and output projection module. Leveraging this structure, we develop customized complex-valued transformers for three representative applications in wireless communications: channel estimation, user activity detection, and joint design of pilot, feedback quantization, and precoder. These applications utilize transformers with varying levels of sophistication and span a variety of tasks, ranging from regression to classification, supervised to unsupervised learning, and specific module design to end-to-end design. Experimental results verify the theoretical advantage and effectiveness of the complex-valued transformers for the above three applications compared to other traditional real-valued neural network-based methods.

A great amount of endeavor has recently been devoted to activity detection for massive machine-type communications in cell-free multiple-input multiple-output (MIMO) systems. However, as the number of antennas at the access points (APs) increases, the Rayleigh distance that separates the near-field and far-field regions also expands, rendering the conventional assumption of sole far-field propagation impractical. To address this challenge, this paper establishes a covariance-based formulation that can effectively capture the statistical property of hybrid near-far field channels. Based on this formulation, we theoretically reveal that increasing the proportion of near-field channels enhances the detection performance. Furthermore, we propose a distributed algorithm, where each AP performs local activity detection and only exchanges the detection results to the central processing unit, thus significantly reduces the computational complexity and the communication overhead. Not only with convergence guarantee, the proposed algorithm is unified in the sense that it can handle single-cell or cell-free systems with either near-field or far-field devices as special cases. Simulation results validate the theoretical analyses and demonstrate the superior performance of the proposed approach compared with existing methods.

Joint activity detection and channel estimation is a crucial task in grant-free random access for massive machine-type communications. To enhance communication quality, intelligent reflecting surfaces (IRSs) have been proposed as a promising technology by controlling the propagation environment with passive reflecting elements. However, due to their passive nature and the non-Gaussian device-IRS-BS composite channels, IRSs introduce significant challenges for joint activity detection and channel estimation. To this end, this paper establishes an accurate statistical model for the composite channel, demonstrating that it follows a variance-gamma (VG) distribution. Based on the exact channel statistics, this paper employs a Bernoulli-VG prior and extends the standard approximate message passing algorithm to learn the Gamma-distributed channel variance within an expectation-maximization framework. Additionally, this paper introduces a novel approach to enforce consistency in device activity status across all base station antennas by transforming the activity detection task into the estimation of a dedicated active probability for each device. Extensive simulations validate the proposed VG channel model and demonstrate significant improvement due to imposing consistent activity probability across multiple antennas.

The recently emerged movable antenna (MA) shows great potential in leveraging spatial degrees of freedom for enhancing the performance of wireless systems. However, resource allocation in MA-aided systems faces unique challenges due to the non-convex and coupled constraints on antenna positions. This paper systematically reveals the challenges brought by the minimum MA separation constraints, and proposes a penalty framework for resource allocation under such new constraints in MA-aided systems. By introducing auxiliary variables, the proposed framework separates the non-convex and coupled antenna distance constraints from the movable region constraint. This enables the resulting problem be efficiently solved by alternating optimization, where the optimization of the original variables resembles that in conventional resource allocation problem while the optimization with respect to the auxiliary variables is achieved in closed-form solutions. To illustrate the effectiveness of the proposed framework, we present three case studies: capacity maximization, latency minimization, and regularized zero-forcing precoding. Simulation results demonstrate that the proposed optimization framework consistently outperforms state-of-the-art schemes.

Joint activity detection and channel estimation present significant challenges in massive machine-type communication under grant-free random access. Due to the intermittency of grant-free random access, these tasks can be framed as a compressive sensing problem, where approximate message passing (AMP) has been a popular algorithm to recover the sparse channels. However, existing AMP-based algorithms typically estimate the channels at each antenna independently, largely ignoring the common sparsity support among channels at different antennas and access points. This oversight leads to inadequate performance of AMP-based activity detection and channel estimation, especially under short pilot sequence lengths. To leverage the common sparsity patterns of multiple unknown channel vectors, this paper proposes a Bayesian model that induces block sparsity through a consistent activity status across all channels associated with each device. By treating the activity probability of each device as an unknown parameter, this paper derives an AMP-based expectation-maximization (EM) algorithm to jointly learn the activity probability and the channels. It is shown theoretically that under mild conditions, when the number of antennas goes to infinity, the recovered active status is guaranteed to be the ground truth. Furthermore, to overcome the limitation of the standard AMP, whose convergence to the state evolution relies on large pilot sequence lengths, another algorithm based on the vector AMP, with its convergence depending on the number of devices rather than pilot sequence length, is also proposed. Simulation results demonstrate that the proposed algorithms require a much shorter pilot length than existing state-of- the-art AMP-based methods to achieve the same performance. Notably, under the same pilot length, the vector AMP-based EM algorithm achieves even higher detection accuracy than the covariance-based method in cell-free systems.

Various adverse weather conditions such as fog and rain pose a significant challenge to autonomous driving (AD) perception tasks like semantic segmentation, object detection, etc. The common domain adaption strategy is to minimize the disparity between images captured in clear and adverse weather conditions. However, domain adaption faces two challenges: (I) it typically relies on utilizing clear image as a reference, which is challenging to obtain in practice; (II) it generally targets single adverse weather condition and performs poorly when confronting the mixture of multiple adverse weather conditions. To address these issues, we introduce a reference-free and Adverse weather condition-independent (Advent) framework (rather than a specific model architecture) that can be implemented by various backbones and heads. This is achieved by leveraging the homogeneity over short durations, getting rid of clear reference and being generalizable to arbitrary weather condition. Specifically, Advent includes three integral components: (I) Locally Sequential Mechanism (LSM) leverages temporal correlations between adjacent frames to achieve the weather-condition-agnostic effect thanks to the homogeneity behind arbitrary weather condition; (II) Globally Shuffled Mechanism (GSM) is proposed to shuffle segments processed by LSM from different positions of input sequence to prevent the overfitting to LSM-induced temporal patterns; (III) Unfolded Regularizers (URs) are the deep unfolding implementation of two proposed regularizers to penalize the model complexity to enhance across-weather generalization. We take the semantic segmentation task as an example to assess the proposed Advent framework. Extensive experiments demonstrate that the proposed Advent outperforms existing state-of-the-art baselines with large margins.

This paper addresses the challenge of large model (LM)-embedded wireless network for handling the trade-off problem of model accuracy and network latency. To guarantee a high-quality of users' service, the network latency should be minimized while maintaining an acceptable inference accuracy. To meet this requirement, LM quantization is proposed to reduce the latency. However, the excessive quantization may destroy the accuracy of LM inference. To this end, a promising fluid antenna (FA) technology is investigated for enhancing the transmission capacity, leading to a lower network latency in the LM-embedded multiple-input multiple-output (MIMO) network. To design the FA-assisted LM-embedded network with the lower latency and higher accuracy requirements, the latency and peak signal to noise ratio (PSNR) are considered in the objective function. Then, an efficient optimization algorithm is proposed under the block coordinate descent framework. Simulation results are provided to show the convergence behavior of the proposed algorithm, and the performance gains from the proposed FA-assisted LMembedded network over the other benchmark networks in terms of network latency and PSNR.

In the realm of activity detection for massive machine-type communications, intelligent reflecting surfaces (IRS) have shown significant potential in enhancing coverage for devices lacking direct connections to the base station (BS). However, traditional activity detection methods are typically designed for a single type of channel model, which does not reflect the complexities of real-world scenarios, particularly in systems incorporating IRS. To address this challenge, this paper introduces a novel approach that combines model-driven deep unfolding with a mixture of experts (MoE) framework. By automatically selecting one of three expert designs and applying it to the unfolded projected gradient method, our approach eliminates the need for prior knowledge of channel types between devices and the BS. Simulation results demonstrate that the proposed MoE-augmented deep unfolding method surpasses the traditional covariance-based method and black-box neural network design, delivering superior detection performance under mixed channel fading conditions.

A great amount of endeavor has recently been devoted to activity detection for massive machine-type communications in cell-free massive MIMO. However, in practice, as the number of antennas at the access points (APs) increases, the Rayleigh distance that separates the near-field and far-field regions also expands, rendering the conventional assumption of far-field propagation alone impractical. To address this challenge, this paper considers a hybrid near-far field activity detection in cell-free massive MIMO, and establishes a covariance-based formulation, which facilitates the development of a distributed algorithm to alleviate the computational burden at the central processing unit (CPU). Specifically, each AP performs local activity detection for the devices and then transmits the detection result to the CPU for further processing. In particular, a novel coordinate descent algorithm based on the Sherman-Morrison-Woodbury update with Taylor expansion is proposed to handle the local detection problem at each AP. Moreover, we theoretically analyze how the hybrid near-far field channels affect the detection performance. Simulation results validate the theoretical analysis and demonstrate the superior performance of the proposed approach compared with existing approaches.

The development of edge computing places critical demands on energy-efficient model deployment for multiple-input multiple-output (MIMO) detection tasks. Deploying deep unfolding models such as PGD-Nets and ADMM-Nets into resource-constrained edge devices using quantization methods is challenging. Existing quantization methods based on quantization aware training (QAT) suffer from performance degradation due to their reliance on parametric distribution assumption of activations and static quantization step sizes. To address these challenges, this paper proposes a novel kernel-based adaptive quantization (KAQ) framework for deep unfolding networks. By utilizing a joint kernel density estimation (KDE) and maximum mean discrepancy (MMD) approach to align activation distributions between full-precision and quantized models, the need for prior distribution assumptions is eliminated. Additionally, a dynamic step size updating method is introduced to adjust the quantization step size based on the channel conditions of wireless networks. Extensive simulations demonstrate that the accuracy of proposed KAQ framework outperforms traditional methods and successfully reduces the model's inference latency.

Device activity detection in the emerging cell-free massive multiple-input multiple-output systems has been recognized as a crucial task in machine-type communications, in which multiple access points jointly identify the active devices from a large number of potential devices based on the received signals. Most of the existing works addressing this problem rely on the impractical assumption that different active devices transmit signals synchronously. However, in practice, synchronization cannot be guaranteed due to the low-cost oscillators, which brings additional discontinuous and nonconvex constraints to the detection problem. To address this challenge, this paper reveals an equivalent reformulation to the asynchronous activity detection problem, which facilitates the development of a distributed algorithm that satisfies the highly nonconvex constraints in a gentle fashion as the iteration number increases. To reduce the capacity requirements of the fronthauls, we further design a communication-efficient accelerated distributed algorithm. Simulation results demonstrate that the proposed two distributed algorithms outperform state-of-the-art approaches. Moreover, the accelerated distributed algorithm requires a very small number of quantization bits to approach the ideal detection performance.

Photoemission plays a central role in a wide range of fields, from electronic structure measurements to free-electron laser sources. In metallic emitters, single-photon 1 , multiphoton 2 – 5 or strong-field emission 6 – 10 processes are the primary photoemission mechanisms. Here, using a sub-work-function 3.06 eV continuous-wave laser, photoemission from waveguide-integrated monolayer graphene is observed to occur at peak power densities >5 orders of magnitude lower than reported multiphoton and strong-field emission 6 , 11 , 12 . The behaviour is explained by the emission of hot electrons in graphene. In monolayer graphene, the need for photoelectrons to be transported to an emitting surface is eliminated, dramatically enhancing the probability of emission before thermalization. These results indicate that integrated-photonics-driven hot-electron emission provides a rich new area of exploration for both electron emission and integrated photonics.

Using hot electrons to drive electrochemical reactions has drawn considerable interest in driving high-barrier reactions and enabling efficient solar to fuel conversion. However, the conversion efficiency from hot electrons to electrochemical products is typically low due to high hot electron scattering rates. Here, it is shown that the hydrogen evolution reaction (HER) in an acidic solution can be efficiently modulated by hot electrons injected into a thin gold film by an Au-Al2O3-Si metal-insulator-semiconductor (MIS) junction. Despite the large scattering rates in gold, it is shown that the hot electron driven HER can reach quantum efficiencies as high as ∼85% with a shift in the onset of hydrogen evolution by ∼0.6 V. By simultaneously measuring the currents from the solution, gold, and silicon terminals during the experiments, we find that the HER rate can be decomposed into three components: (i) thermal electron, corresponding to the thermal electron distribution in gold; (ii) hot electron, corresponding to electrons injected from silicon into gold which drive the HER before fully thermalizing; and (iii) silicon direct injection, corresponding to electrons injected from Si into gold that drive the HER before electron-electron scattering occurs. Through a series of control experiments, we eliminate the possibility of the observed HER rate modulation coming from lateral resistivity of the thin gold film, pinholes in the gold, oxidation of the MIS device, and measurement circuit artifacts. Next, we theoretically evaluate the feasibility of hot electron injection modifying the available supply of electrons. Considering electron-electron and electron-phonon scattering, we track how hot electrons injected at different energies interact with the gold-solution interface as they scatter and thermalize. The simulator is first used to reproduce other published experimental pump-probe hot electron measurements, and then simulate the experimental conditions used here. These simulations predict that hot electron injection first increases the supply of electrons to the gold-solution interface at higher energies by several orders of magnitude and causes a peaked electron interaction with the gold-solution interface at the electron injection energy. The first prediction corresponds to the observed hot electron electrochemical current, while the second prediction corresponds to the observed silicon direct injection current. These results indicate that MIS devices offer a versatile platform for hot electron sources that can efficiently drive electrochemical reactions.

As Moore's law comes to an end, the search for additional integrated circuit functionality has shifted from scaling down lateral dimensions to combining multiple materials onto a single substrate. However, the quality of crystalline semiconductors is highly sensitive to the substrate upon which it is grown, preventing multiple materials from being directly grown on single substrates. To circumvent this challenge, many complex growth strategies have been developed, such as strain relaxation buffer-layers, nanostructure growth, and template selective epitaxy. However, even with these advanced growth techniques, the growth of manufacturable crystalline materials is still limited to crystalline surfaces. Here, the authors show that using templated liquid phase (TLP) growth, single crystalline indium phosphide on Si can be grown using a variety of buffers, both crystalline and amorphous. Moreover, by performing detailed optoelectronic characterization, the authors find that the quality of the grown material not only closely matches commercial single crystalline InP wafers but is also highly insensitive to the buffer layer used. This unique feature of TLP growth could enable the next generation of crystalline material integration.

The growth of crystalline compound semiconductors on amorphous and non-epitaxial substrates is a fundamental challenge for state-of-the-art thin-film epitaxial growth techniques. Direct growth of materials on technologically relevant amorphous surfaces, such as nitrides or oxides results in nanocrystalline thin films or nanowire-type structures, preventing growth and integration of high-performance devices and circuits on these surfaces. Here, we show crystalline compound semiconductors grown directly on technologically relevant amorphous and non-epitaxial substrates in geometries compatible with standard microfabrication technology. Furthermore, by removing the traditional epitaxial constraint, we demonstrate an atomically sharp lateral heterojunction between indium phosphide and tin phosphide, two materials with vastly different crystal structures, a structure that cannot be grown with standard vapor-phase growth approaches. Critically, this approach enables the growth and manufacturing of crystalline materials without requiring a nearly lattice-matched substrate, potentially impacting a wide range of fields, including electronics, photonics, and energy devices.

Neuromorphic or "brain-like" computation is a leading candidate for efficient, fault-tolerant processing of large-scale data as well as real-time sensing and transduction of complex multivariate systems and networks such as self-driving vehicles or Internet of Things applications. In biology, the synapse serves as an active memory unit in the neural system and is the component responsible for learning and memory. Electronically emulating this element via a compact, scalable technology which can be integrated in a three-dimensional (3-D) architecture is critical for future implementations of neuromorphic processors. However, present day 3-D transistor implementations of synapses are typically based on low-mobility semiconductor channels or technologies that are not scalable. Here, we demonstrate a crystalline indium phosphide (InP)-based artificial synapse for spiking neural networks that exhibits elasticity, short-term plasticity, long-term plasticity, metaplasticity, and spike timing-dependent plasticity, emulating the critical behaviors exhibited by biological synapses. Critically, we show that this crystalline InP device can be directly integrated via back-end processing on a Si wafer using a SiO2 buffer without the need for a crystalline seed, enabling neuromorphic devices that can be implemented in a scalable and 3-D architecture. Specifically, the device is a crystalline InP channel field-effect transistor that interacts with neuron spikes by modification of the population of filled traps in the MOS structure itself. Unlike other transistor-based implementations, we show that it is possible to mimic these biological functions without the use of external factors (e.g., surface adsorption of gas molecules) and without the need for the high electric fields necessary for traditional flash-based implementations. Finally, when exposed to neuronal spikes with a waveform similar to that observed in the brain, these devices exhibit the ability to learn without the need for any external potentiating/depressing circuits, mimicking the biological process of Hebbian learning.

The authors report the experimental demonstration of independent control over work function and field enhancement factor in hybrid field emitters using a lanthanum hexaboride (LaB6) nanoparticle low-work function coating on monolayer graphene on microfabricated silicon arrays. A critical challenge in field emitters is combining the scalability and uniformity of silicon microfabrication with low-work function materials. Specifically, the authors engineer the field enhancement through microfabrication of the underlying silicon wafers and control the work function by the transfer and deposition of monolayer graphene and LaB6 nanoparticles. Using this coating, the turn-on electric field, defined as the electric field required for 10 μA/cm2 of emission current density, drops by 5× from 12.5 to 2.6 V/μm. To further analyze these results, the authors carried out detailed electronic and structural characterization of the hybrid emitters to experimentally determine the work function and model the field enhancement factor of the physical structure. Using these coupled simulations and experiments, the authors show that the work function and field enhancement factor can be independently controlled, potentially enabling ultralow turn on, uniform, and stable emitters.

Recent developments in nanophotonics have provided a clear roadmap for improving the efficiency of photonic devices through control over absorption and emission of devices. These advances could prove transformative for a wide variety of devices, such as photovoltaics, photoelectrochemical devices, photodetectors, and light-emitting diodes. However, it is often challenging to physically create the nanophotonic designs required to engineer the optical properties of devices. Here, we present a platform based on crystalline indium phosphide that enables thin-film nanophotonic structures with physical morphologies that are impossible to achieve through conventional state-of-the-art material growth techniques. Here, nanostructured InP thin films have been demonstrated on non-epitaxial alumina inverted nanocone (i-cone) substrates via a low-cost and scalable thin-film vapor-liquid-solid growth technique. In this process, indium films are first evaporated onto the i-cone structures in the desired morphology, followed by a high-temperature step that causes a phase transformation of the indium into indium phosphide, preserving the original morphology of the deposited indium. Through this approach, a wide variety of nanostructured film morphologies are accessible using only control over evaporation process variables. Critically, the as-grown nanotextured InP thin films demonstrate excellent optoelectronic properties, suggesting this platform is promising for future high-performance nanophotonic devices.