Searcharxiv⌕ Search

arXiv subjects

Xuehe Wang

Publications and source records attributed to Xuehe Wang.

At least 19 recordsLinked to original sources
arXiv2026

Beyond Memory Majority: Latent-Source Reasoning for Multi-Agent Memory Arbitration

Long-term multi-agent systems continuously accumulate the memories produced by different agents. Existing memory methods typically treat retrieved memories as independent evidence and combine them through voting or weighting. However, this independence assumption often fails in multi-agent settings: memories written by different agents may inherit the same upstream source or shared bias, causing correlated evidence to be repeatedly counted and creating a false majority. We term this failure mode \textit{Memory Correlation Bias}. To address the issue, we propose the \textbf{C}orrelation-\textbf{A}ware \textbf{M}emory \textbf{A}rbitration (CAMA) framework that jointly decouples retrieved memories and recovers missing independent evidence. We model the retrieved memories as query-conditioned evidence groups and combine neural dependency inference with provenance-based symbolic priors to estimate the effective number of independent evidence sources, thereby preventing correlated memories from forming a false majority. Since critical independent evidence may be absent from the initial retrieval set, \textsc{CAMA} further learns a sequential recovery policy that actively retrieves alternative evidence or traces upstream sources before making the final decision, aiming to recover sufficient independent evidence for reliable arbitration while minimizing retrieval cost. Experiments on multiple benchmarks demonstrate the superiority of our method over the state-of-the-art baseline methods, suppressing false majorities induced by correlated memories.

cs.AI↗
arXiv2026

FEDBUD: Joint Incentive and Privacy Optimization for Resource-Constrained Federated Learning

Federated learning has become a popular paradigm for privacy protection and edge-based machine learning. However, defending against differential attacks and devising incentive strategies remain significant bottlenecks in this field. Despite recent works on privacy-aware incentive mechanism design for federated learning, few of them consider both data volume and noise level. In this paper, we propose a novel federated learning system called FEDBUD, which combines privacy and economic concerns together by considering the joint influence of data volume and noise level on incentive strategy determination. In this system, the cloud server controls monetary payments to edge nodes, while edge nodes control data volume and noise level that potentially impact the model performance of the cloud server. To determine the mutually optimal strategies for both sides, we model FEDBUD as a two-stage Stackelberg Game and derive the Nash Equilibrium using the mean-field estimator and virtual queue. Experimental results on real-world datasets demonstrate the outstanding performance of FEDBUD.

cs.DC↗
arXiv2026

Verify Before You Commit: Towards Faithful Reasoning in LLM Agents via Self-Auditing

In large language model (LLM) agents, reasoning trajectories are treated as reliable internal beliefs for guiding actions and updating memory. However, coherent reasoning can still violate logical or evidential constraints, allowing unsupported beliefs repeatedly stored and propagated across decision steps, leading to systematic behavioral drift in long-horizon agentic systems. Most existing strategies rely on the consensus mechanism, conflating agreement with faithfulness. In this paper, inspired by the vulnerability of unfaithful intermediate reasoning trajectories, we propose \textbf{S}elf-\textbf{A}udited \textbf{Ve}rified \textbf{R}easoning (\textsc{SAVeR}), a novel framework that enforces verification over internal belief states within the agent before action commitment, achieving faithful reasoning. Concretely, we structurally generate persona-based diverse candidate beliefs for selection under a faithfulness-relevant structure space. To achieve reasoning faithfulness, we perform adversarial auditing to localize violations and repair through constraint-guided minimal interventions under verifiable acceptance criteria. Extensive experiments on six benchmark datasets demonstrate that our approach consistently improves reasoning faithfulness while preserving competitive end-task performance.

cs.AI↗
arXiv2026

TianQuan-S2S: A Subseasonal-to-Seasonal Global Weather Model via Incorporate Climatology State

Accurate Subseasonal-to-Seasonal (S2S) forecasting is vital for decision-making in agriculture, energy production, and emergency management. However, it remains a challenging and underexplored problem due to the chaotic nature of the weather system. Recent data-driven studies have shown promising results, but their performance is limited by the inadequate incorporation of climate states and a model tendency to degrade, progressively losing fine-scale details and yielding over-smoothed forecasts. To overcome these limitations, we propose TianQuan-S2S, a global S2S forecasting model that integrates initial weather states with climatological means via incorporating climatology into patch embedding and enhancing variability capture through an uncertainty-augmented Transformer. Extensive experiments on the Earth Reanalysis 5 (ERA5) reanalysis dataset demonstrate that our model yields a significant improvement in both deterministic and ensemble forecasting over the climatology mean, traditional numerical methods, and data-driven models. Ablation studies empirically show the effectiveness of our model designs. Remarkably, our model outperforms skillful numerical ECMWF-S2S and advanced data-driven Fuxi-S2S in key meteorological variables. The code implementation can be found in https://github.com/zhangminglang42/TianQuan.

cs.LG↗
arXiv2026

Chain-of-Context Learning: Dynamic Constraint Understanding for Multi-Task VRPs

Multi-task Vehicle Routing Problems (VRPs) aim to minimize routing costs while satisfying diverse constraints. Existing solvers typically adopt a unified reinforcement learning (RL) framework to learn generalizable patterns across tasks. However, they often overlook the constraint and node dynamics during the decision process, making the model fail to accurately react to the current context. To address this limitation, we propose Chain-of-Context Learning (CCL), a novel framework that progressively captures the evolving context to guide fine-grained node adaptation. Specifically, CCL constructs step-wise contextual information via a Relevance-Guided Context Reformulation (RGCR) module, which adaptively prioritizes salient constraints. This context then guides node updates through a Trajectory-Shared Node Re-embedding (TSNR) module, which aggregates shared node features from all trajectories' contexts and uses them to update inputs for the next step. By modeling evolving preferences of the RL agent, CCL captures step-by-step dependencies in sequential decision-making. We evaluate CCL on 48 diverse VRP variants, including 16 in-distribution and 32 out-of-distribution (with unseen constraints) tasks. Experimental results show that CCL performs favorably against the state-of-the-art baselines, achieving the best performance on all in-distribution tasks and the majority of out-of-distribution tasks.

cs.AI↗
arXiv2025

Task-Adaptive Parameter-Efficient Fine-Tuning for Weather Foundation Models

While recent advances in machine learning have equipped Weather Foundation Models (WFMs) with substantial generalization capabilities across diverse downstream tasks, the escalating computational requirements associated with their expanding scale increasingly hinder practical deployment. Current Parameter-Efficient Fine-Tuning (PEFT) methods, designed for vision or language tasks, fail to address the unique challenges of weather downstream tasks, such as variable heterogeneity, resolution diversity, and spatiotemporal coverage variations, leading to suboptimal performance when applied to WFMs. To bridge this gap, we introduce WeatherPEFT, a novel PEFT framework for WFMs incorporating two synergistic innovations. First, during the forward pass, Task-Adaptive Dynamic Prompting (TADP) dynamically injects the embedding weights within the encoder to the input tokens of the pre-trained backbone via internal and external pattern extraction, enabling context-aware feature recalibration for specific downstream tasks. Furthermore, during backpropagation, Stochastic Fisher-Guided Adaptive Selection (SFAS) not only leverages Fisher information to identify and update the most task-critical parameters, thereby preserving invariant pre-trained knowledge, but also introduces randomness to stabilize the selection. We demonstrate the effectiveness and efficiency of WeatherPEFT on three downstream tasks, where existing PEFT methods show significant gaps versus Full-Tuning, and WeatherPEFT achieves performance parity with Full-Tuning using fewer trainable parameters. The code of this work is available at https://github.com/ShileiCao/WeatherPEFT.

cs.LG↗
arXiv2025

Degree of Staleness-Aware Data Updating in Federated Learning

Handling data staleness remains a significant challenge in federated learning with highly time-sensitive tasks, where data is generated continuously and data staleness largely affects model performance. Although recent works attempt to optimize data staleness by determining local data update frequency or client selection strategy, none of them explore taking both data staleness and data volume into consideration. In this paper, we propose DUFL(Data Updating in Federated Learning), an incentive mechanism featuring an innovative local data update scheme manipulated by three knobs: the server's payment, outdated data conservation rate, and clients' fresh data collection volume, to coordinate staleness and volume of local data for best utilities. To this end, we introduce a novel metric called DoS(the Degree of Staleness) to quantify data staleness and conduct a theoretic analysis illustrating the quantitative relationship between DoS and model performance. We model DUFL as a two-stage Stackelberg game with dynamic constraint, deriving the optimal local data update strategy for each client in closed-form and the approximately optimal strategy for the server. Experimental results on real-world datasets demonstrate the significant performance of our approach.

cs.LG↗
arXiv2025

Multi-Hop Privacy Propagation for Differentially Private Federated Learning in Social Networks

Federated learning (FL) enables collaborative model training across decentralized clients without sharing local data, thereby enhancing privacy and facilitating collaboration among clients connected via social networks. However, these social connections introduce privacy externalities: a client's privacy loss depends not only on its privacy protection strategy but also on the privacy decisions of others, propagated through the network via multi-hop interactions. In this work, we propose a socially-aware privacy-preserving FL mechanism that systematically quantifies indirect privacy leakage through a multi-hop propagation model. We formulate the server-client interaction as a two-stage Stackelberg game, where the server, as the leader, optimizes incentive policies, and clients, as followers, strategically select their privacy budgets, which determine their privacy-preserving levels by controlling the magnitude of added noise. To mitigate information asymmetry in networked privacy estimation, we introduce a mean-field estimator to approximate the average external privacy risk. We theoretically prove the existence and convergence of the fixed point of the mean-field estimator and derive closed-form expressions for the Stackelberg Nash Equilibrium. Despite being designed from a client-centric incentive perspective, our mechanism achieves approximately-optimal social welfare, as revealed by Price of Anarchy (PoA) analysis. Experiments on diverse datasets demonstrate that our approach significantly improves client utilities and reduces server costs while maintaining model performance, outperforming both Social-Agnostic (SA) baselines and methods that account for social externalities.

cs.LG↗
arXiv2024

A Game-Theoretic Framework for Privacy-Aware Client Sampling in Federated Learning

This paper aims to design a Privacy-aware Client Sampling framework in Federated learning, named FedPCS, to tackle the heterogeneous client sampling issues and improve model performance. First, we obtain a pioneering upper bound for the accuracy loss of the FL model with privacy-aware client sampling probabilities. Based on this, we model the interactions between the central server and participating clients as a two-stage Stackelberg game. In Stage I, the central server designs the optimal time-dependent reward for cost minimization by considering the trade-off between the accuracy loss of the FL model and the rewards allocated. In Stage II, each client determines the correction factor that dynamically adjusts its privacy budget based on the reward allocated to maximize its utility. To surmount the obstacle of approximating other clients' private information, we introduce the mean-field estimator to estimate the average privacy budget. We analytically demonstrate the existence and convergence of the fixed point for the mean-field estimator and derive the Stackelberg Nash Equilibrium to obtain the optimal strategy profile. By rigorously theoretical convergence analysis, we guarantee the robustness of FedPCS. Moreover, considering the conventional sampling strategy in privacy-preserving FL, we prove that the random sampling approach's PoA can be arbitrarily large. To remedy such efficiency loss, we show that the proposed privacy-aware client sampling strategy successfully reduces PoA, which is upper bounded by a reachable constant. To address the challenge of varying privacy requirements throughout different training phases in FL, we extend our model and analysis and derive the adaptive optimal sampling ratio for the central server. Experimental results on different datasets demonstrate the superiority of FedPCS compared with the existing SOTA FL strategies under IID and Non-IID datasets.

cs.GT↗
arXiv2024

FedAgg: Adaptive Federated Learning with Aggregated Gradients

Federated Learning (FL) has emerged as a crucial distributed training paradigm, enabling discrete devices to collaboratively train a shared model under the coordination of a central server, while leveraging their locally stored private data. Nonetheless, the non-independent-and-identically-distributed (Non-IID) data generated on heterogeneous clients and the incessant information exchange among participants may significantly impede training efficacy, retard the model convergence rate and increase the risk of privacy leakage. To alleviate the divergence between the local and average model parameters and obtain a fast model convergence rate, we propose an adaptive FEDerated learning algorithm called FedAgg by refining the conventional stochastic gradient descent (SGD) methodology with an AGgregated Gradient term at each local training epoch and adaptively adjusting the learning rate based on a penalty term that quantifies the local model deviation. To tackle the challenge of information exchange among clients during local training and design a decentralized adaptive learning rate for each client, we introduce two mean-field terms to approximate the average local parameters and gradients over time. Through rigorous theoretical analysis, we demonstrate the existence and convergence of the mean-field terms and provide a robust upper bound on the convergence of our proposed algorithm. The extensive experimental results on real-world datasets substantiate the superiority of our framework in comparison with existing state-of-the-art FL strategies for enhancing model performance and accelerating convergence rate under IID and Non-IID datasets.

cs.LG↗
arXiv2024

Adaptive Federated Learning via New Entropy Approach

Federated Learning (FL) has emerged as a prominent distributed machine learning framework that enables geographically discrete clients to train a global model collaboratively while preserving their privacy-sensitive data. However, due to the non-independent-and-identically-distributed (Non-IID) data generated by heterogeneous clients, the performances of the conventional federated optimization schemes such as FedAvg and its variants deteriorate, requiring the design to adaptively adjust specific model parameters to alleviate the negative influence of heterogeneity. In this paper, by leveraging entropy as a new metric for assessing the degree of system disorder, we propose an adaptive FEDerated learning algorithm based on ENTropy theory (FedEnt) to alleviate the parameter deviation among heterogeneous clients and achieve fast convergence. Nevertheless, given the data disparity and parameter deviation of heterogeneous clients, determining the optimal dynamic learning rate for each client becomes a challenging task as there is no communication among participating clients during the local training epochs. To enable a decentralized learning rate for each participating client, we first introduce the mean-field terms to estimate the components associated with other clients' local parameters. Furthermore, we provide rigorous theoretical analysis on the existence and determination of the mean-field estimators. Based on the mean-field estimators, the closed-form adaptive learning rate for each client is derived by constructing the Hamilton equation. Moreover, the convergence rate of our proposed FedEnt is proved. The extensive experimental results on the real-world datasets (i.e., MNIST, EMNIST-L, CIFAR10, and CIFAR100) show that our FedEnt algorithm surpasses FedAvg and its variants (i.e., FedAdam, FedProx, and FedDyn) under Non-IID settings and achieves a faster convergence rate.

cs.DC↗
arXiv2024

QI-DPFL: Quality-Aware and Incentive-Boosted Federated Learning with Differential Privacy

Federated Learning (FL) has increasingly been recognized as an innovative and secure distributed model training paradigm, aiming to coordinate multiple edge clients to collaboratively train a shared model without uploading their private datasets. The challenge of encouraging mobile edge devices to participate zealously in FL model training procedures, while mitigating the privacy leakage risks during wireless transmission, remains comparatively unexplored so far. In this paper, we propose a novel approach, named QI-DPFL (Quality-Aware and Incentive-Boosted Federated Learning with Differential Privacy), to address the aforementioned intractable issue. To select clients with high-quality datasets, we first propose a quality-aware client selection mechanism based on the Earth Mover's Distance (EMD) metric. Furthermore, to attract high-quality data contributors, we design an incentive-boosted mechanism that constructs the interactions between the central server and the selected clients as a two-stage Stackelberg game, where the central server designs the time-dependent reward to minimize its cost by considering the trade-off between accuracy loss and total reward allocated, and each selected client decides the privacy budget to maximize its utility. The Nash Equilibrium of the Stackelberg game is derived to find the optimal solution in each global iteration. The extensive experimental results on different real-world datasets demonstrate the effectiveness of our proposed FL framework, by realizing the goal of privacy protection and incentive compatibility.

cs.GT↗
arXiv2022

Incentive Mechanism and Path Planning for UAV Hitching over Traffic Networks

Package delivery via the UAVs is a promising transport mode to provide efficient and green logistic services, especially in urban areas or complicated topography. However, the energy storage limit of the UAV makes it difficult to perform long-distance delivery tasks. In this paper, we propose a novel multimodal logistics framework, in which the UAVs can call on ground vehicles to provide hitch services to save their own energy and extend their delivery distance. This multimodal logistics framework is formulated as a two-stage model to jointly consider the incentive mechanism design for ground vehicles and path planning for UAVs. In Stage I, to deal with the motivations for ground vehicles to assist UAV delivery, a dynamic pricing scheme is proposed to best balance the vehicle response time and payments to ground vehicles. It shows that a higher price should be decided if the vehicle response time is long to encourage more vehicles to offer a ride. In Stage II, the task allocation and path planning of the UAVs over traffic network is studied based on the vehicle response time obtained in Stage I. To address pathfinding with restrictions and the performance degradation of the pathfinding algorithm due to the rising number of conflicts in multi-agent pathfinding, we propose the suboptimal conflict avoidance-based path search (CABPS) algorithm, which has polynomial time complexity. Finally, we validate our results via simulations. It is shown that our approach is able to increase the success rate of UAV package delivery. Moreover, we estimate the delivery time of the UAV in a pessimistic case, it is still twice as fast as the delivery time of the ground vehicle only.

eess.SY↗
arXiv2022

Dynamic Pricing for Client Recruitment in Federated Learning

Though federated learning (FL) well preserves clients' data privacy, many clients are still reluctant to join FL given the communication cost and energy consumption in their mobile devices. It is important to design pricing compensations to motivate enough clients to join FL and distributively train the global model. Prior pricing mechanisms for FL are static and cannot adapt to clients' random arrival pattern over time. We propose a new dynamic pricing solution in closed-form by constructing the Hamiltonian function to optimally balance the client recruitment time and the model training time, without knowing clients' actual arrivals or training costs. During the client recruitment phase, we offer time-dependent monetary rewards per client arrival to trade-off between the total payment and the FL model's accuracy loss. Such reward gradually increases when we approach to the recruitment deadline or have greater data aging, and we also extend the deadline if the clients' training time per iteration becomes shorter. Further, we extend to consider heterogeneous client types in training data size and training time per iteration. We successfully extend our dynamic pricing solution and develop an optimal algorithm of linear complexity to monotonically select client types for FL. Finally, we also show the robustness of our solution against estimation error of clients' data sizes and run numerical experiments to validate our conclusion.

cs.GT↗
arXiv2021

Dynamic Pricing and Mean Field Analysis for Controlling Age of Information

Today many mobile users in various zones are invited to sense and send back real-time useful information (e.g., traffic observation and sensor data) to keep the freshness of the content updates in such zones. However, due to the sampling cost in sensing and transmission, a user may not have the incentive to contribute the real-time information to help reduce the age of information (AoI). We propose dynamic pricing for each zone to offer age-dependent monetary returns and encourage users to sample information at different rates over time. This dynamic pricing design problem needs to well balance the monetary payments as rewards to users and the AoI evolution over time, and is challenging to solve especially under the incomplete information about users' arrivals and their private sampling costs. After formulating the problem as a nonlinear constrained dynamic program, to avoid the curse of dimensionality, we first propose to approximate the dynamic AoI reduction as a time-average term and successfully solve the approximate dynamic pricing in closed-form. Further, by providing the steady-state analysis for an infinite time horizon, we show that the pricing scheme (though in closed-form) can be further simplified to an $\varepsilon$-optimal version without recursive computing over time. Finally, we extend the AoI control from a single zone to many zones with heterogeneous user arrival rates and initial ages, where each zone cares not only its own AoI dynamics but also the average AoI of all the zones in a mean field game system to provide a holistic service. Accordingly, we propose decentralized mean field pricing for each zone to self-operate by using a mean field term to estimate the average age dynamics of all the zones, which does not even require many zones to exchange their local data with each other.

eess.SY↗
arXiv2019

Dynamic Pricing for Controlling Age of Information

Fueled by the rapid development of communication networks and sensors in portable devices, today many mobile users are invited by content providers to sense and send back real-time useful information (e.g., traffic observations and sensor data) to keep the freshness of the providers' content updates. However, due to the sampling cost in sensing and transmission, an individual may not have the incentive to contribute the real-time information to help a content provider reduce the age of information (AoI). Accordingly, we propose dynamic pricing for the provider to offer age-dependent monetary returns and encourage users to sample information at different rates over time. This dynamic pricing design problem needs to balance the monetary payments to users and the AoI evolution over time, and is challenging to solve especially under the incomplete information about users' arrivals and their private sampling costs. For analysis tractability, we linearize the nonlinear AoI evolution in the constrained dynamic programming problem, by approximating the dynamic AoI reduction as a time-average term and solving the approximate dynamic pricing in closed-form. Then, we estimate this approximate term based on Brouwer's fixed-point theorem. Finally, we provide the steady-state analysis of the optimized approximate dynamic pricing scheme for an infinite time horizon, and show that the pricing scheme can be further simplified to an e-optimal version without recursive computing over time.

cs.IT↗
arXiv2019

Economic Analysis of Rollover and Shared Data Plans

In today's growing data market, wireless service providers (WSPs) compete severely to attract users by announcing innovative data plans. Two of the most popular innovative data plans are rollover and shared data plans, where the former plan allows a user to keep his unused data quota to next month and the latter plan allows users in a family to share unused data. As a pioneer to provide such data plans, a WSP faces immediate revenue loss from existing users who pay less overage charges due to less data over-usage, but his market share increases gradually by attracting new users and those under the other WSPs. In some countries, WSPs have asymmetric timing for providing such innovative data plans, while some other markets' WSPs have symmetric timing or no planning. This raises the question of why and when the competitive WSPs should offer the new data plans. This paper provides game theoretic modelling and analysis of the WSPs' timing of offering innovative data plans, by considering new user arrival and dynamic user churn between WSPs. Our equilibrium analysis shows that the WSP with small market share prefers to announce the innovative data plan first to attract more users, while the WSP with large market share prefers to announce later to avoid the immediate revenue loss. In a market with many new users, WSPs with similar market shares will offer the data plans simultaneously, but these WSPs facing few new users may not offer any new plan. Perhaps surprisingly, WSPs' profits can decrease with new user number and they may not benefit from the option of innovative data plans. Finally, unlike rollover data plan, we show that the timing of shared data plan further depends on the composition of users.

cs.NI↗
arXiv2018

Dynamic Pricing and Capacity Allocation of UAV-provided Mobile Services

Due to its agility and mobility, the unmanned aerial vehicle (UAV) is a promising technology to provide high-quality mobile services (e.g., fast Internet access, edge computing, and local caching) to ground users. Major Internet Service Providers (ISPs) want to enable UAV-provided services (UPS) to improve and enrich the current mobile services for additional profit. This profit-maximization problem is not easy as the UAV has limited energy storage and needs to fly closely to serve users, requiring an optimal energy allocation for balancing both hovering time and service capacity. When hovering in a hotspot, how the UAV should dynamically price its capacity-limited UPS according to randomly arriving users with private service valuations is another question. We prove that the UAV should ask for a higher price if the leftover hovering time is longer or its service capacity is smaller, and its expected profit approaches to that under complete user information if the hovering time is sufficiently large. As the hotspot's user occurrence rate increases, a shorter hovering time or a larger service capacity should be allocated. Finally, when the UAV faces multiple hotspot candidates with different user occurrence rates and flying distances, we prove that it is optimal to deploy the UAV to serve a single hotspot. With multiple UAVs, however, this result can be reversed with UAVs' forking deployment to different hotspots.

eess.SP↗