The evaluation of Generative AI (GenAI) systems plays a critical role in public policy and decision-making, yet existing methods are often limited by reliance on benchmark-driven, point-estimate comparisons that fail to capture uncertainty and broader societal impacts. This paper argues for the use of Bayesian statistics as a principled framework to address these challenges. Bayesian methods enable the integration of domain expertise through prior elicitation, allow for continuous learning from new data, and provide robust uncertainty quantification via posterior inference. We demonstrate how Bayesian inference can be applied to GenAI evaluation, particularly in incorporating stakeholder perspectives to enhance fairness, transparency, and reliability. Furthermore, we discuss Bayesian workflows as an iterative process for model validation and refinement, ensuring robust assessments of GenAI systems in dynamic, real-world contexts.
This work considers the three-dimensional (3-D) positioning problem in a Terahertz (THz) system enabled by a modular extra-large (XL) array with sub-connected architecture. Our purpose is to estimate the Cartesian Coordinates of multiple user equipments (UEs) with the received signal of the RF chains while considering the spatial non-stationarity (SNS). We apply the hybrid spherical-planar wave model (HSPWM) as the channel model owing to the structual feature of the modular array, and propose a 3-D localization algorithm with relatively high accuracy and low complexity. Specifically, we first distinguish the visible sub-arrays (SAs) located in the VR and estimate the angles-of-arrival (AoAs) from each UE to typical visible SAs with the largest receive power via compressed sensing (CS) method. In addition, we apply the weighted least square (WLS) method to obtain a coarse 3-D position estimation of each UE according to the AoA estimations. Then, we estimate the AoAs of the other SAs with a reduced dictionary (RD)-CS-based method for lower computational complexity, and utilize all the efficient AoA estimations to derive a fine position estimation. Simulation results indicate that the proposed positioning framework based on modular XL-array can achieve satisfactory accuracy with evident reduction in complexity. Furthermore, the deployment of SAs and the allocation of antenna elements need to be specially designed for better positioning performance.
In addition to enhancing wireless communication coverage quality, reconfigurable intelligent surface (RIS) technique can also assist in positioning. In this work, we consider RIS-assisted superimposed pilot and data transmission without the assumption availability of prior channel state information and position information of mobile user equipments (UEs). To tackle this challenge, we design a frame structure of transmission protocol composed of several location coherence intervals, each with pure-pilot and data-pilot transmission durations. The former is used to estimate UE locations, while the latter is time-slotted, duration of which does not exceed the channel coherence time, where the data and pilot signals are transmitted simultaneously. We conduct the Fisher Information matrix (FIM) analysis and derive \text {Cram\'er-Rao bound} (CRB) for the position estimation error. The inverse fast Fourier transform (IFFT) is adopted to obtain the estimation results of UE positions, which are then exploited for channel estimation. Furthermore, we derive the closed-form lower bound of the ergodic achievable rate of superimposed pilot (SP) transmission, which is used to optimize the phase profile of the RIS to maximize the achievable sum rate using the genetic algorithm. Finally, numerical results validate the accuracy of the UE position estimation using the IFFT algorithm and the superiority of the proposed SP scheme by comparison with the regular pilot scheme.
The rise of reinforcement learning (RL) in critical real-world applications demands a fundamental rethinking of privacy in AI systems. Traditional privacy frameworks, designed to protect isolated data points, fall short for sequential decision-making systems where sensitive information emerges from temporal patterns, behavioral strategies, and collaborative dynamics. Modern RL paradigms, such as federated RL (FedRL) and RL with human feedback (RLHF) in large language models (LLMs), exacerbate these challenges by introducing complex, interactive, and context-dependent learning environments that traditional methods do not address. In this position paper, we argue for a new privacy paradigm built on four core principles: multi-scale protection, behavioral pattern protection, collaborative privacy preservation, and context-aware adaptation. These principles expose inherent tensions between privacy, utility, and interpretability that must be navigated as RL systems become more pervasive in high-stakes domains like healthcare, autonomous vehicles, and decision support systems powered by LLMs. To tackle these challenges, we call for the development of new theoretical frameworks, practical mechanisms, and rigorous evaluation methodologies that collectively enable effective privacy protection in sequential decision-making systems.
In recent years, image compression for high-level vision tasks has attracted considerable attention from researchers. Given that object information in images plays a far more crucial role in downstream tasks than background information, some studies have proposed semantically structuring the bitstream to selectively transmit and reconstruct only the information required by these tasks. However, such methods structure the bitstream after encoding, meaning that the coding process still relies on the entire image, even though much of the encoded information will not be transmitted. This leads to redundant computations. Traditional image compression methods require a two-dimensional image as input, and even if the unimportant regions of the image are set to zero by applying a semantic mask, these regions still participate in subsequent computations as part of the image. To address such limitations, we propose an image compression method based on a position-indexed self-attention mechanism that encodes and decodes only the visible parts of the masked image. Compared to existing semantic-structured compression methods, our approach can significantly reduce computational costs.
Contemporary digital technology has a pivotal role in the design of customized medical appliances, including occlusal splints used in the treatment of stomatognathic system dysfunctions. We present an approach to computer-aided design and precision assessment of positioning occlusal splints, bridging clinical concepts with current digital dental practice. In our model, a 3D splint is generated based on a transformation matrix that represents the therapeutic change in mandibular position, defined by a specialist using a virtual patient model reconstructed from intraoral scans, CBCT, 3D facial scans and plaster model digitisation. The paper introduces a novel method for generating splints that accurately reproduce occlusal conditions in the therapeutic position, including a mechanism for resolving surface conflicts through virtual embossing. We demonstrate how transformation matrices can be acquired through clinical tools and intraoral devices, and evaluate the accuracy of the designed and printed splints using profile and surface deviation analysis. The proposed method enables reproducible, patient-specific splint fabrication and opens new possibilities in diagnostics, multimodal image registration and quantification of occlusal discrepancies.
This technical report introduces a targeted improvement to the StreamPETR framework, specifically aimed at enhancing velocity estimation, a critical factor influencing the overall NuScenes Detection Score. While StreamPETR exhibits strong 3D bounding box detection performance as reflected by its high mean Average Precision our analysis identified velocity estimation as a substantial bottleneck when evaluated on the NuScenes dataset. To overcome this limitation, we propose a customized positional embedding strategy tailored to enhance temporal modeling capabilities. Experimental evaluations conducted on the NuScenes test set demonstrate that our improved approach achieves a state-of-the-art NDS of 70.86% using the ViT-L backbone, setting a new benchmark for camera-only 3D object detection.
A control design modification to prevent integrator windup for position saturated servo-controllers is introduced. The design is based on the formalism of Control Barrier Functions and represents an anti-windup integrator modification for position-limited servo-controllers. The method is applicable to Linear Time Invariant Multi-Input-Multi-Output open-loop stable continuous time systems. A flight control application example of the developed anti-windup control solution is discussed.
Many works have succeeded in reconstructing Gaussian human avatars from multi-view videos. However, they either struggle to capture pose-dependent appearance details with a single MLP, or rely on a computationally intensive neural network to reconstruct high-fidelity appearance but with rendering performance degraded to non-real-time. We propose a novel Gaussian human avatar representation that can reconstruct high-fidelity pose-dependence appearance with details and meanwhile can be rendered in real time. Our Gaussian avatar is empowered by spatially distributed MLPs which are explicitly located on different positions on human body. The parameters stored in each Gaussian are obtained by interpolating from the outputs of its nearby MLPs based on their distances. To avoid undesired smooth Gaussian property changing during interpolation, for each Gaussian we define a set of Gaussian offset basis, and a linear combination of basis represents the Gaussian property offsets relative to the neutral properties. Then we propose to let the MLPs output a set of coefficients corresponding to the basis. In this way, although Gaussian coefficients are derived from interpolation and change smoothly, the Gaussian offset basis is learned freely without constraints. The smoothly varying coefficients combined with freely learned basis can still produce distinctly different Gaussian property offsets, allowing the ability to learn high-frequency spatial signals. We further use control points to constrain the Gaussians distributed on a surface layer rather than allowing them to be irregularly distributed inside the body, to help the human avatar generalize better when animated under novel poses. Compared to the state-of-the-art method, our method achieves better appearance quality with finer details while the rendering speed is significantly faster under novel views and novel poses.
Mobile robots on construction sites require accurate pose estimation to perform autonomous surveying and inspection missions. Localization in construction sites is a particularly challenging problem due to the presence of repetitive features such as flat plastered walls and perceptual aliasing due to apartments with similar layouts inter and intra floors. In this paper, we focus on the global re-positioning of a robot with respect to an accurate scanned mesh of the building solely using LiDAR data. In our approach, a neural network is trained on synthetic LiDAR point clouds generated by simulating a LiDAR in an accurate real-life large-scale mesh. We train a diffusion model with a PointNet++ backbone, which allows us to model multiple position candidates from a single LiDAR point cloud. The resulting model can successfully predict the global position of LiDAR in confined and complex sites despite the adverse effects of perceptual aliasing. The learned distribution of potential global positions can provide multi-modal position distribution. We evaluate our approach across five real-world datasets and show the place recognition accuracy of 77% +/-2m on average while outperforming baselines at a factor of 2 in mean error.
Training a state-of-the-art Large Language Model (LLM) is an increasingly expensive endeavor due to growing computational, hardware, energy, and engineering demands. Yet, an often-overlooked (and seldom paid) expense is the human labor behind these models' training data. Every LLM is built on an unfathomable amount of human effort: trillions of carefully written words sourced from books, academic papers, codebases, social media, and more. This position paper aims to assign a monetary value to this labor and argues that the most expensive part of producing an LLM should be the compensation provided to training data producers for their work. To support this position, we study 64 LLMs released between 2016 and 2024, estimating what it would cost to pay people to produce their training datasets from scratch. Even under highly conservative estimates of wage rates, the costs of these models' training datasets are 10-1000 times larger than the costs to train the models themselves, representing a significant financial liability for LLM providers. In the face of the massive gap between the value of training data and the lack of compensation for its creation, we highlight and discuss research directions that could enable fairer practices in the future.
In robot-assisted laparoscopic radical prostatectomy (RALP), the location of the instrument tip is important to register the ultrasound frame with the laparoscopic camera frame. A long-standing limitation is that the instrument tip position obtained from the da Vinci API is inaccurate and requires hand-eye calibration. Thus, directly computing the position of the tool tip in the camera frame using the vision-based method becomes an attractive solution. Besides, surgical instrument tip detection is the key component of other tasks, like surgical skill assessment and surgery automation. However, this task is challenging due to the small size of the tool tip and the articulation of the surgical instrument. Surgical instrument segmentation becomes relatively easy due to the emergence of the Segmentation Foundation Model, i.e., Segment Anything. Based on this advancement, we explore the deep learning-based surgical instrument tip detection approach that takes the part-level instrument segmentation mask as input. Comparison experiments with a hand-crafted image-processing approach demonstrate the superiority of the proposed method on simulated and real datasets.
In multi-target tracking and detection tasks, it is necessary to continuously track multiple targets, such as vehicles, pedestrians, etc. To achieve this goal, the system must be able to continuously acquire and process image frames containing these targets. These consecutive frame images enable the algorithm to update the position and state of the target in real-time in each frame of the image. How to accurately associate the detected target with the target in the previous or next frame to form a stable trajectory is a complex problem. Therefore, a multi object tracking and detection method for intelligent driving vehicles based on YOLOv5 and point cloud 3D projection is proposed. Using Retinex algorithm to enhance the image of the environment in front of the vehicle, remove light interference in the image, and build an intelligent detection model based on YOLOv5 network structure. The enhanced image is input into the model, and multiple targets in front of the vehicle are identified through feature extraction and target localization. By combining point cloud 3D projection technology, the correlation between the position changes of adjacent frame images in the projection coordinate system can be inferred. By sequentially projecting the multi-target recognition results of multiple consecutive frame images into the 3D laser point cloud environment, effective tracking of the motion trajectories of all targets in front of the vehicle can be achieved. The experimental results show that the application of this method for intelligent driving vehicle front multi-target tracking and detection yields a MOTA (Tracking Accuracy) value greater than 30, demonstrating its superior tracking and detection performance.
Floor plans can provide valuable prior information that helps enhance the accuracy of indoor positioning systems. However, existing research typically faces challenges in efficiently leveraging floor plan information and applying it to complex indoor layouts. To fully exploit information from floor plans for positioning, we propose a floor plan-assisted fusion positioning algorithm (FP-BP) using Bluetooth low energy (BLE) and pedestrian dead reckoning (PDR). In the considered system, a user holding a smartphone walks through a positioning area with BLE beacons installed on the ceiling, and can locate himself in real time. In particular, FP-BP consists of two phases. In the offline phase, FP-BP programmatically extracts map features from a stylized floor plan based on their binary masks, and constructs a mapping function to identify the corresponding map feature of any given position on the map. In the online phase, FP-BP continuously computes BLE positions and PDR results from BLE signals and smartphone sensors, where a novel grid-based maximum likelihood estimation (GML) algorithm is introduced to enhance BLE positioning. Then, a particle filter is used to fuse them and obtain an initial estimate. Finally, FP-BP performs post-position correction to obtain the final position based on its specific map feature. Experimental results show that FP-BP can achieve a real-time mean positioning accuracy of 1.19 m, representing an improvement of over 28% compared to existing floor plan-fused baseline algorithms.
In the context of the joint radar and communications (JRC) framework, reconfigurable intelligent surfaces (RISs) emerged as a promising technology for their ability to shape the propagation environment by adjusting their phase-shift coefficients. However, achieving perfect synchronization and effective collaboration between access points (APs) and RISs is crucial to successful operation. This paper investigates the performance of a bistatic JRC network operating in the millimeter-wave (mmWave) frequency band, where the receiving AP is equipped with an RIS-integrated array. This system simultaneously serves multiple UEs while estimating the position of a target with limited prior knowledge of its position. To achieve this, we optimize both the power allocation of the transmitted waveform and the RIS phase-shift matrix to minimize the position error bound (PEB) of the target. At the same time, we ensure that the UEs achieve an acceptable level of spectral efficiency. The numerical results show that an RIS-integrated array, even with a small number of receiving antennas, can achieve high localization accuracy. Additionally, optimized phase-shifts significantly improve the localization accuracy in comparison to a random phase-shift configuration.
Recent advances in reconfigurable antennas have led to the new concept of the fluid antenna system (FAS) for shape and position flexibility, as another degree of freedom for wireless communication enhancement. This paper explores the integration of a transmit FAS array for hybrid beamforming (HBF) into a reconfigurable intelligent surface (RIS)-assisted communication architecture for multiuser communications in the downlink, corresponding to the downlink RIS-assisted multiuser multiple-input single-output (MISO) FAS model (Tx RIS-assisted-MISO-FAS). By considering Rician channel fading, we formulate a sum-rate maximization optimization problem to alternately optimize the HBF matrix, the RIS phase-shift matrix, and the FAS position. Due to the strong coupling of multiple optimization variables, the multi-fractional summation in the sum-rate expression, the modulus-1 limitation of analog phase shifters and RIS, and the antenna position variables appearing in the exponent, this problem is highly non-convex, which is addressed through the block coordinate descent (BCD) framework in conjunction with semidefinite relaxation (SDR) and majorization-minimization (MM) methods. To reduce the computational complexity, we then propose a low-complexity grating-lobe (GL)-based telescopic-FAS (TFA) with multiple delicately deployed RISs under the sub-connected HBF architecture and the line-of-sight (LoS)-dominant channel condition, to allow closed-form solutions for the HBF and TFA position. Our simulation results illustrate that the former optimization scheme significantly enhances the achievable rate of the proposed system, while the GL-based TFA scheme also provides a considerable gain over conventional fixed-position antenna (FPA) systems, requiring statistical channel state information (CSI) only and with low computational complexity.
A Pinching-Antenna SyStem (PASS)-assisted convert communication framework is proposed. PASS utilizes dielectric waveguides with freely positioned pinching antennas (PAs) to establish strong line-of-sight links. Capitalizing on this high reconfigurable flexibility of antennas, the potential of PASS for covert communications is investigated. 1)~For the single-waveguide single-PA (SWSP) scenario, a closed-form optimal PA position that maximizes the covert rate is first derived. Subsequently, a one-dimensional power search is employed to enable low-complexity optimization for covert communications. With antenna mobility on a scale of meters, PASS can deal with the challenging situation of the eavesdropper enjoying better channel conditions than the legal user. 2)~For the multi-waveguide multi-PA (MWMP) scenario, the positions of multiple PAs are optimized to enable effective pinching beamforming, thereby enhancing the covert rate. To address the resultant multimodal joint transmit and pinching beamforming problem, a twin particle swarm optimization (TwinPSO) approach is proposed. Numerical results demonstrate that: i)~the proposed approaches can effectively resolve the optimization problems; ii)~PASS achieves a higher covert rate than conventional fixed-position antenna architectures; and iii)~with enhanced flexibility, the MWMP setup outperforms the SWSP counterpart.
We investigate joint localization and synchronization in the downlink of a distributed multiple-input-multiple-output (D-MIMO) system, aiming to estimate the position and phase offset of a single-antenna user equipment (UE) using downlink transmissions of multiple phase-synchronized, multi-antenna access points (APs). We propose two transmission protocols: sequential (P1) and simultaneous (P2) AP transmissions, together with the ML estimators that either leverage (coherent estimator) or disregard phase information (non-coherent estimator). Simulation results reveal that downlink D-MIMO holds significant potential for high-accuracy localization while showing that P2 provides superior localization performance and reduced transmission latency.
Integrated sensing and communication (ISAC) has emerged as a key enabler for next-generation wireless networks, supporting advanced applications such as high-precision localization and environment reconstruction. Cooperative ISAC (CoISAC) further enhances these capabilities by enabling multiple base stations (BSs) to jointly optimize communication and sensing performance through coordination. However, CoISAC beamforming design faces significant challenges due to system heterogeneity, large-scale problem complexity, and sensitivity to parameter estimation errors. Traditional deep learning-based techniques fail to exploit the unique structural characteristics of CoISAC systems, thereby limiting their ability to enhance system performance. To address these challenges, we propose a Link-Heterogeneous Graph Neural Network (LHGNN) for joint beamforming in CoISAC systems. Unlike conventional approaches, LHGNN models communication and sensing links as heterogeneous nodes and their interactions as edges, enabling the capture of the heterogeneous nature and intricate interactions of CoISAC systems. Furthermore, a graph attention mechanism is incorporated to dynamically adjust node and link importance, improving robustness to channel and position estimation errors. Numerical results demonstrate that the proposed attention-enhanced LHGNN achieves superior communication rates while maintaining sensing accuracy under power constraints. The proposed method also exhibits strong robustness to communication channel and position estimation error.
Gallice and Monzon (2019) present a natural environment that sustains full cooperation in one-shot social dilemmas among a finite number of self-interested agents. They demonstrate that in a sequential public goods game, where agents lack knowledge of their position in the sequence but can observe some predecessors' actions, full contribution emerges in equilibrium due to agents' incentive to induce potential successors to follow suit. Furthermore, they show that this principle extends to a number of social dilemmas, with the prominent example that of the prisoner's dilemma. In this study, we experimentally test the theoretical predictions of this model in a multi- player prisoner's dilemma environment, where subjects are not aware of their position in the sequence and receive only partial information on past cooperating actions. We test the predictions of the model, and through rigorous structural econometric analysis, we test the descriptive capacity of the model against alternative behavioural strategies, such as conditional cooperation, altruistic play and free-riding behaviour. We find that the majority resorts to free-riding behaviour, around 30% is classified as Gallice and Monzon (2019) types, followed by those with social preference considerations and the unconditional altruists.