arXiv subjects

Ji Wang

Publications and source records attributed to Ji Wang.

At least 91 records · Page 5Linked to original sources

arXiv2025

Energy-Efficient Design for Downlink Pinching-Antenna Systems with QoS Guarantee

Pinching antennas have recently garnered significant attention due to their ability to dynamically reconfigure wireless propagation environments. Despite notable advancements in this area, the exploration of energy efficiency (EE) maximization in pinching-antenna systems remains relatively underdeveloped. In this paper, we address the EE maximization problem in a downlink time-division multiple access (TDMA)-based multi-user system employing one waveguide and multiple pinching antennas, where each user is subject to a minimum rate constraint to ensure quality-of-service. The formulated optimization problem jointly considers transmit power and time allocations as well as the positioning of pinching antennas, resulting in a non-convex problem. To tackle this challenge, we first obtain the optimal positions of the pinching antennas. Based on this, we establish a feasibility condition for the system. Subsequently, the joint power and time allocation problem is decomposed into two subproblems, which are solved iteratively until convergence. Specifically, the power allocation subproblem is addressed through an iterative approach, where a semi-analytical solution is obtained in each iteration. Likewise, a semi-analytical solution is derived for the time allocation subproblem. Numerical simulations demonstrate that the proposed pinching-antenna-based strategy significantly outperforms both conventional fixed-antenna systems and other benchmark pinching-antenna schemes in terms of EE.

cs.IT↗

arXiv2025

Energy-Efficient Resource Allocation for NOMA-Assisted Uplink Pinching-Antenna Systems

The pinching-antenna architecture has emerged as a promising solution for reconfiguring wireless propagation environments and enhancing system performance. While prior research has primarily focused on sum-rate maximization or transmit power minimization of pinching-antenna systems, the critical aspect of energy efficiency (EE) has received limited attention. Given the increasing importance of EE in future wireless communication networks, this work investigates EE optimization in a non-orthogonal multiple access (NOMA)-assisted multi-user pinching-antenna uplink system. The problem entails the joint optimization of the users' transmit power and the pinching-antenna position. The resulting optimization problem is non-convex due to tightly coupled variables. To tackle this, we employ an alternating optimization framework to decompose the original problem into two subproblems: one focusing on power allocation and the other on antenna positioning. A low-complexity optimal solution is derived for the power allocation subproblem, while the pinching-antenna positioning subproblem is addressed using a particle swarm optimization algorithm to obtain a high-quality near-optimal solution. Simulation results demonstrate that the proposed scheme significantly outperforms both conventional-antenna configurations and orthogonal multiple access-based pinching-antenna systems in terms of EE.

cs.IT↗

arXiv2025

Channel Estimation for Pinching-Antenna Systems (PASS)

Pinching Antennas (PAs) represent a revolutionary flexible antenna technology that leverages dielectric waveguides and electromagnetic coupling to mitigate large-scale path loss. This letter is the first to explore channel estimation for Pinching-Antenna SyStems (PASS), addressing their uniquely ill-conditioned and underdetermined channel characteristics. In particular, two efficient deep learning-based channel estimators are proposed. 1) PAMoE: This estimator incorporates dynamic padding, feature embedding, fusion, and mixture of experts (MoE) modules, which effectively leverage the positional information of PAs and exploit expert diversity. 2) PAformer: This Transformer-style estimator employs the self-attention mechanism to predict channel coefficients in a per-antenna manner, which offers more flexibility to adaptively deal with dynamic numbers of PAs in practical deployment. Numerical results demonstrate that 1) the proposed deep learning-based channel estimators outperform conventional methods and exhibit excellent zero-shot learning capabilities, and 2) PAMoE delivers higher channel estimation accuracy via MoE specialization, while PAformer natively handles an arbitrary number of PAs, trading self-attention complexity for superior scalability.

cs.IT↗

arXiv2025

SacFL: Self-Adaptive Federated Continual Learning for Resource-Constrained End Devices

The proliferation of end devices has led to a distributed computing paradigm, wherein on-device machine learning models continuously process diverse data generated by these devices. The dynamic nature of this data, characterized by continuous changes or data drift, poses significant challenges for on-device models. To address this issue, continual learning (CL) is proposed, enabling machine learning models to incrementally update their knowledge and mitigate catastrophic forgetting. However, the traditional centralized approach to CL is unsuitable for end devices due to privacy and data volume concerns. In this context, federated continual learning (FCL) emerges as a promising solution, preserving user data locally while enhancing models through collaborative updates. Aiming at the challenges of limited storage resources for CL, poor autonomy in task shift detection, and difficulty in coping with new adversarial tasks in FCL scenario, we propose a novel FCL framework named SacFL. SacFL employs an Encoder-Decoder architecture to separate task-robust and task-sensitive components, significantly reducing storage demands by retaining lightweight task-sensitive components for resource-constrained end devices. Moreover, $\rm{SacFL}$ leverages contrastive learning to introduce an autonomous data shift detection mechanism, enabling it to discern whether a new task has emerged and whether it is a benign task. This capability ultimately allows the device to autonomously trigger CL or attack defense strategy without additional information, which is more practical for end devices. Comprehensive experiments conducted on multiple text and image datasets, such as Cifar100 and THUCNews, have validated the effectiveness of $\rm{SacFL}$ in both class-incremental and domain-incremental scenarios. Furthermore, a demo system has been developed to verify its practicality.

cs.LG↗

arXiv2025

Pinching-Antenna Assisted Simultaneous Wireless Information and Power Transfer

This letter introduces a novel pinching-antenna-system (PASS) assisted simultaneous wireless information and power transfer (SWIPT), where multiple pinching antennas (PAs) are strategically activiated on a waveguide to facilitate information transmission to multiple information receivers (IRs) and power transfer to multiple energy receivers (ERs) simultaneously. Leveraging the single-waveguide architecture, non-orthogonal multiple access is employed to enable the superposed transmission of information signals, eliminating the need for dedicated energy carriers by concurrently serving both the IRs and the ERs. In this letter, an optimization problem is first formulated with an objective is to maximize the sum power received at the ERs via joint optimizing the IRs power allocation and the PAs positioning. Given the challenging non-convex nature of the formulated optimization problem, it is decoupled into two sub-problems which are solved alternatively. Specifically, the power allocation subproblem is recast a convex optimization problem, and hence can be solved efficiently. Furthermore, we propose a high-precision element-wise algorithm and a low-complexity linearly decreasing weight particle swarm optimization algorithm to solve the position optimization sub-problem. The numerical results demonstrate that PAs assisted SWIPT can achieve a remarkable performance gain compared to conventional system.

eess.SP↗

arXiv2025

Efficient Multi-Task Modeling through Automated Fusion of Trained Models

Although multi-task learning is widely applied in intelligent services, traditional multi-task modeling methods often require customized designs based on specific task combinations, resulting in a cumbersome modeling process. Inspired by the rapid development and excellent performance of single-task models, this paper proposes an efficient multi-task modeling method that can automatically fuse trained single-task models with different structures and tasks to form a multi-task model. As a general framework, this method allows modelers to simply prepare trained models for the required tasks, simplifying the modeling process while fully utilizing the knowledge contained in the trained models. This eliminates the need for excessive focus on task relationships and model structure design. To achieve this goal, we consider the structural differences among various trained models and employ model decomposition techniques to hierarchically decompose them into multiple operable model components. Furthermore, we have designed an Adaptive Knowledge Fusion (AKF) module based on Transformer, which adaptively integrates intra-task and inter-task knowledge based on model components. Through the proposed method, we achieve efficient and automated construction of multi-task models, and its effectiveness is verified through extensive experiments on three datasets.

cs.LG↗

arXiv2025

CUT: Pruning Pre-Trained Multi-Task Models into Compact Models for Edge Devices

Multi-task learning has garnered widespread attention in the industry due to its efficient data utilization and strong generalization capabilities, making it particularly suitable for providing high-quality intelligent services to users. Edge devices, as the primary platforms directly serving users, play a crucial role in delivering multi-task services. However, current multi-task models are often large, and user task demands are increasingly diverse. Deploying such models directly on edge devices not only increases the burden on these devices but also leads to task redundancy. To address this issue, this paper innovatively proposes a pre-trained multi-task model pruning method specifically designed for edge computing. The goal is to utilize existing pre-trained multi-task models to construct a compact multi-task model that meets the needs of edge devices. The specific implementation steps are as follows: First, decompose the tasks within the pre-trained multi-task model and select tasks based on actual user needs. Next, while retaining the knowledge of the original pre-trained model, evaluate parameter importance and use a parameter fusion method to effectively integrate shared parameters among tasks. Finally, obtain a compact multi-task model suitable for edge devices. To validate the effectiveness of the proposed method, we conducted experiments on three public image datasets. The experimental results fully demonstrate the superiority and efficiency of this method, providing a new solution for multi-task learning on edge devices.

cs.LG↗

arXiv2025

Multi-task Federated Learning with Encoder-Decoder Structure: Enabling Collaborative Learning Across Different Tasks

Federated learning has been extensively studied and applied due to its ability to ensure data security in distributed environments while building better models. However, clients participating in federated learning still face limitations, as clients with different structures or tasks cannot participate in learning together. In view of this, constructing a federated learning framework that allows collaboration between clients with different model structures and performing different tasks, enabling them to share valuable knowledge to enhance model efficiency, holds significant practical implications for the widespread application of federated learning. To achieve this goal, we propose a multi-task federated learning with encoder-decoder structure (M-Fed). Specifically, given the widespread adoption of the encoder-decoder architecture in current models, we leverage this structure to share intra-task knowledge through traditional federated learning methods and extract general knowledge from the encoder to achieve cross-task knowledge sharing. The training process is similar to traditional federated learning, and we incorporate local decoder and global decoder information into the loss function. The local decoder iteratively updates and gradually approaches the global decoder until sufficient cross-task knowledge sharing is achieved. Our method is lightweight and modular, demonstrating innovation compared to previous research. It enables clients performing different tasks to share general knowledge while maintaining the efficiency of traditional federated learning systems. We conducted experiments on two widely used benchmark datasets to verify the feasibility of M-Fed and compared it with traditional methods. The experimental results demonstrate the effectiveness of M-Fed in multi-task federated learning.

cs.LG↗

arXiv2025

DeepONet of dynamic event-triggered backstepping boundary control for reaction-diffusion PDEs

We present an event-triggered boundary control scheme for a class of reaction-diffusion PDEs using operator learning and backstepping method. Our first-of-its-kind contribution aims at learning the backstepping kernels, which inherently induces the learning of the gains in the event trigger and the control law. The kernel functions in constructing the control law are approximated with neural operators (NOs) to improve the computational efficiency. Then, a dynamic event-triggering mechanism is designed, based on the plant and the continuous-in-time control law using kernels given by NOs,to determine the updating times of the actuation signal. In the resulting event-based closed-loop system, a strictly positive lower bound of the minimal dwell time is found, which is independent of initial conditions. As a result, the absence of a Zeno behavior is guaranteed. Besides, exponential convergence to zero of the L_2 norm of the reaction-diffusion PDE state and the dynamic variable in the event-triggering mechanism is proved via Lyapunov analysis. The effectiveness of the proposed method is illustrated by numerical simulation.

math.OC↗

arXiv2025

Kepler's Platonic Model and Its Application to Exoplanetary Systems

Johannes Kepler's attempt to explain the arrangement of the six innermost planets of the Solar System using his Platonic Solid Model-which postulates that planetary orbits are nested within the five Platonic solids-was ultimately unsuccessful. However, while his model failed to describe our own planetary system, Kepler was remarkably prescient in hypothesizing the existence of exoplanetary systems that might conform to this geometric framework. In this study, we analyze all known multiple exoplanet systems containing three to six planets and identify those that best match the Keplerian Platonic model. Using a semi-major-axis (SMA) ratio metric defined as the sum of squared differences between observed and theoretical semi-major-axis ratios, we find that the most well-matched three-, four-, five-, and six-planet exoplanetary systems exhibit significantly lower discrepancy values $(4.38 \times 10^{-6}, 1.05 \times 10^{-2}, 8.21 \times 10^{-2}$, and $2.43 \times 10^{-1}$, respectively) compared to the inner six planets of the Solar System at 12.68. These results demonstrate that Kepler's Platonic Model is applicable to certain exoplanetary systems, suggesting that while the Solar System does not adhere to this idealized structure, other planetary systems may be governed by underlying geometric and mathematical principles akin to Kepler's vision. This study highlights the special nature of these exoplanetary systems and their potential alignment with the Platonic five-element framework.

astro-ph.EP↗

arXiv2025

Unlocking Generalization Power in LiDAR Point Cloud Registration

In real-world environments, a LiDAR point cloud registration method with robust generalization capabilities (across varying distances and datasets) is crucial for ensuring safety in autonomous driving and other LiDAR-based applications. However, current methods fall short in achieving this level of generalization. To address these limitations, we propose UGP, a pruned framework designed to enhance generalization power for LiDAR point cloud registration. The core insight in UGP is the elimination of cross-attention mechanisms to improve generalization, allowing the network to concentrate on intra-frame feature extraction. Additionally, we introduce a progressive self-attention module to reduce ambiguity in large-scale scenes and integrate Bird's Eye View (BEV) features to incorporate semantic information about scene elements. Together, these enhancements significantly boost the network's generalization performance. We validated our approach through various generalization experiments in multiple outdoor scenes. In cross-distance generalization experiments on KITTI and nuScenes, UGP achieved state-of-the-art mean Registration Recall rates of 94.5% and 91.4%, respectively. In cross-dataset generalization from nuScenes to KITTI, UGP achieved a state-of-the-art mean Registration Recall of 90.9%. Code will be available at https://github.com/peakpang/UGP.

cs.CV↗

arXiv2025

PEPSI Investigation, Retrieval, and Atlas of Numerous Giant Atmospheres (PIRANGA). I. The Ubiquity of Fe I Emission and Inversions in Ultra Hot Jupiter Atmospheres

We present high-resolution optical emission spectroscopy observations of the ultra hot Jupiters (UHJs) TOI-1431 b and TOI-1518 b using the PEPSI spectrograph on the LBT. We detect emission lines from Fe I with a significance of 5.68 $σ$ and 7.68 $σ$ for TOI 1431 b and TOI-1518 b, respectively. We also tentatively detect Cr I emission from TOI-1431 b at $4.32σ$. For TOI-1518 b, we tentatively detect Ni I, Fe II, and Mg I at significance levels ranging from $3-4σ$. Detection of emission lines indicates that both planets possess temperature inversions in their atmospheres, providing further evidence of the ubiquity of stratospheres among UHJs. By analyzing the population of hot Jupiters, we compare models that predict the distribution of planets in the temperature-gravity space, and find a recent global circulation model suite from Roth et al. (2024) provides a reasonable match to the observed onset of inversions at $T_{\mathrm{eq}}\sim2000$ K. The ubiquity of strong Fe I emission lines among UHJs, together with the paucity of detections of TiO, suggest that atomic iron is the dominant optical opacity source in their atmospheres and can be responsible for the inversions.

astro-ph.EP↗

arXiv2025

HBTP: Heuristic Behavior Tree Planning with Large Language Model Reasoning

Behavior Trees (BTs) are increasingly becoming a popular control structure in robotics due to their modularity, reactivity, and robustness. In terms of BT generation methods, BT planning shows promise for generating reliable BTs. However, the scalability of BT planning is often constrained by prolonged planning times in complex scenarios, largely due to a lack of domain knowledge. In contrast, pre-trained Large Language Models (LLMs) have demonstrated task reasoning capabilities across various domains, though the correctness and safety of their planning remain uncertain. This paper proposes integrating BT planning with LLM reasoning, introducing Heuristic Behavior Tree Planning (HBTP)-a reliable and efficient framework for BT generation. The key idea in HBTP is to leverage LLMs for task-specific reasoning to generate a heuristic path, which BT planning can then follow to expand efficiently. We first introduce the heuristic BT expansion process, along with two heuristic variants designed for optimal planning and satisficing planning, respectively. Then, we propose methods to address the inaccuracies of LLM reasoning, including action space pruning and reflective feedback, to further enhance both reasoning accuracy and planning efficiency. Experiments demonstrate the theoretical bounds of HBTP, and results from four datasets confirm its practical effectiveness in everyday service robot applications.

cs.RO↗

arXiv2025

Water dissociation and rotational broadening in the atmosphere of KELT-20 b from high-resolution spectroscopy

We present atmospheric retrievals from Keck/KPIC phase II observations of the ultra-hot Jupiter KELT-20/MASCARA-2~b. Previous free retrievals of molecular abundances for ultra-hot Jupiters have been impacted by significant model biases due to variations in vertical abundance profiles, which we address by including molecular dissociation into our retrieval framework as an additional free parameter. We measure the abundance of CO ($\rm \log CO_{MMR} = -2.5^{+0.6}_{-0.5}$) and obtain a lower limit on the abundance of H$_2$O ($\rm \log H{_2}O_{MMR} = -1.5^{+0.8}_{-1.0}$, $>-3.0$ at 95\% confidence) in the atmosphere of \keltb. These abundances yield an atmospheric $\rm C/O = 0.1^{+0.4}_{-0.1}$ ($\rm C/O < 0.9$ at 95\% confidence) and suggest a metallicity approximately solar to $10\times$ solar. H$_2$O is dissociated at pressures below $\log P_{\rm H_2O} = -1.2^{+0.5}_{-0.7}$ bar, roughly consistent with predictions from chemical equilibrium models, and suggesting that the retrieved composition is not a result of assumptions about the vertical mixing profiles. We also constrain the rotational velocity of \keltb\ to $v\sin i = 7.5\pm0.7$ \kms, suggesting the presence of a jet comparable to the sound speed in the direction of the planet's rotation, assuming the actual rotation of the planet is tidally locked.

astro-ph.EP↗

arXiv2025

Unlearning through Knowledge Overwriting: Reversible Federated Unlearning via Selective Sparse Adapter

Federated Learning is a promising paradigm for privacy-preserving collaborative model training. In practice, it is essential not only to continuously train the model to acquire new knowledge but also to guarantee old knowledge the right to be forgotten (i.e., federated unlearning), especially for privacy-sensitive information or harmful knowledge. However, current federated unlearning methods face several challenges, including indiscriminate unlearning of cross-client knowledge, irreversibility of unlearning, and significant unlearning costs. To this end, we propose a method named FUSED, which first identifies critical layers by analyzing each layer's sensitivity to knowledge and constructs sparse unlearning adapters for sensitive ones. Then, the adapters are trained without altering the original parameters, overwriting the unlearning knowledge with the remaining knowledge. This knowledge overwriting process enables FUSED to mitigate the effects of indiscriminate unlearning. Moreover, the introduction of independent adapters makes unlearning reversible and significantly reduces the unlearning costs. Finally, extensive experiments on three datasets across various unlearning scenarios demonstrate that FUSED's effectiveness is comparable to Retraining, surpassing all other baselines while greatly reducing unlearning costs.

cs.LG↗

arXiv2025

MRBTP: Efficient Multi-Robot Behavior Tree Planning and Collaboration

Multi-robot task planning and collaboration are critical challenges in robotics. While Behavior Trees (BTs) have been established as a popular control architecture and are plannable for a single robot, the development of effective multi-robot BT planning algorithms remains challenging due to the complexity of coordinating diverse action spaces. We propose the Multi-Robot Behavior Tree Planning (MRBTP) algorithm, with theoretical guarantees of both soundness and completeness. MRBTP features cross-tree expansion to coordinate heterogeneous actions across different BTs to achieve the team's goal. For homogeneous actions, we retain backup structures among BTs to ensure robustness and prevent redundant execution through intention sharing. While MRBTP is capable of generating BTs for both homogeneous and heterogeneous robot teams, its efficiency can be further improved. We then propose an optional plugin for MRBTP when Large Language Models (LLMs) are available to reason goal-related actions for each robot. These relevant actions can be pre-planned to form long-horizon subtrees, significantly enhancing the planning speed and collaboration efficiency of MRBTP. We evaluate our algorithm in warehouse management and everyday service scenarios. Results demonstrate MRBTP's robustness and execution efficiency under varying settings, as well as the ability of the pre-trained LLM to generate effective task-specific subtrees for MRBTP.

cs.RO↗

arXiv2025

Early Accretion of Large Amounts of Solids for Directly-Imaged Exoplanets

As the number of planetary mass objects (PMOs, $\lessapprox$13 M$_{\rm{Jupiter}}$) at wider separation ($\gtrapprox$10 AU) grows, there is emerging evidence that they form differently from their higher-mass brown-dwarf (BD) counterparts. Specifically, PMOs' atmospheres are often enriched by metals and show a large dispersion of metallicity, which is usually interpreted as a sign of solid accretion. {As a first step toward a population-level study of the amount and timing of solid accretion, }we analyze a sample of seven directly-imaged exoplanets with measured stellar and planetary chemical abundances (51 Eri b, $β$ Pic b, HIP 65426 b, HR 8799 c and e, AF Lep b, and YSES 1 c). Our analysis uses existing data of stellar and planetary atmospheric metallicities, and adopts a Bayesian framework that marginalizes the probabilities of disk conditions, formation locations, {planetary interior structures}, and accretion physics. We show that these PMOs accrete large amounts of solids {regardless of whether they form via core accretion or disk instability}. On average $\gtrapprox$50 M$_\oplus$ solids are accreted to enrich planet atmospheres. {Individual planet accretes between 23.3 and 223.2 M$_\oplus$ of solid mass, more than 75\% of which is assumed to stay in the atmosphere and increase the observed metallicity.} The result implies that the solid accretion process and therefore the planet formation process {likely take place} at an early stage {($\lessapprox$2 Myr)} when large amounts of solids are available in young {massive} protoplanetary disks.

astro-ph.EP↗

arXiv2025

OphthBench: A Comprehensive Benchmark for Evaluating Large Language Models in Chinese Ophthalmology

Large language models (LLMs) have shown significant promise across various medical applications, with ophthalmology being a notable area of focus. Many ophthalmic tasks have shown substantial improvement through the integration of LLMs. However, before these models can be widely adopted in clinical practice, evaluating their capabilities and identifying their limitations is crucial. To address this research gap and support the real-world application of LLMs, we introduce the OphthBench, a specialized benchmark designed to assess LLM performance within the context of Chinese ophthalmic practices. This benchmark systematically divides a typical ophthalmic clinical workflow into five key scenarios: Education, Triage, Diagnosis, Treatment, and Prognosis. For each scenario, we developed multiple tasks featuring diverse question types, resulting in a comprehensive benchmark comprising 9 tasks and 591 questions. This comprehensive framework allows for a thorough assessment of LLMs' capabilities and provides insights into their practical application in Chinese ophthalmology. Using this benchmark, we conducted extensive experiments and analyzed the results from 39 popular LLMs. Our evaluation highlights the current gap between LLM development and its practical utility in clinical settings, providing a clear direction for future advancements. By bridging this gap, we aim to unlock the potential of LLMs and advance their development in ophthalmology.

cs.CL↗