Wenhao Yuan, Chenchen Lin, Jian Chen, Jinfeng Xu, Xuehe Wang, Edith Cheuk-Han Ngai. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Federated crowdsourcing has emerged as a promising paradigm for collaboratively solving learning tasks over mobile edge devices, enabling decentralized model training without direct data uploading. However, conventional federated crowdsourcing systems largely overlook two fundamental challenges: (i) the lack of effective incentive mechanisms under unknown participant quality, (ii) the risk of privacy leakage from model uploading. In this paper, we propose an incentive mechanism for federated crowdsourcing based on the Stackelberg game considering$\rho$-zero-concentrated differential privacy andCombinatorialMulti-ArmedBandit mechanism, calledFedCMAB, to tackle the client selection problem with unknown quality and incentive design while preserving confidential information. We model the interaction between the server and clients as a Stackelberg game, where the server dynamically selects a subset of clients and designs rewards to minimize the cost while each client strategically determines its privacy budget to maximize its individual utility. We theoretically establish the existence of the Stackelberg Nash equilibrium. Next, we leverage the combinatorial multi-armed bandit (CMAB) method to learn optimal client selection strategies with provable regret guarantees and derive an upper bound on the cumulative regret of the proposed mechanism. Moreover, we conduct a rigorous Price of Anarchy (PoA) analysis to quantify the efficiency gap between the decentralized equilibrium induced by self-interested clients and the socially optimal solution. Our analyses demonstrate that conventional incentive-agnostic strategy can lead to an unbounded PoA, resulting in severe efficiency loss. In contrast, ourFedCMABframework provably bounds the PoA by a finite constant. Extensive experimental results on multiple datasets demonstrate thatFedCMABconsistently outperforms state-of-the-art baseline methods on both independent and identically distributed (IID) and non-IID data.

Federated Crowdsourcing is a promising and efficient computing paradigm for solving complex tasks on mobile edge devices, while the lack of incentive and the risks of privacy leakage in conventional crowdsourcing platforms have been relatively underexplored so far. In this paper, we propose an incentive mechanism for federated crowdsourcing based on the Stackelberg game considering$\underline{\rho}$-zero-concentrated differential privacy and Combinatorial Multi-Armed Bandit mechanism, called FedCMAB, to tackle the problem of quality unknown client selection and incentive strategy design while preserving participants' confidential information. By selecting a set of clients to maximize the overall training quality, the framework determines an optimal strategy profile in which each participant chooses a privacy budget to maximize their individual utility, while the central server minimizes its cost simultaneously. Through theoretical analysis, we prove the existence of the Stackelberg-Nash Equilibrium and the worst regret. The experimental results on different datasets validate the superiority of our framework.

In the resource constrained unmanned aerial vehicle (UAV) assisted communication networks, a major challenge is how to achieve efficient multi-UAV cooperation with minimal communication overhead. This paper proposes a teammate modeling based resource allocation scheme for a multi-UAV network, where multiple co-channel UAVs cooperatively provide the downlink communication services to the ground users (GUs) with temporal-correlated task demands. We formulate the cooperative resource allocation as a decentralized partially observable Markov decision process (Dec-POMDP), in which the UAVs jointly optimize their trajectory planning, power allocation, and user association policies to maximize the system's cumulative achievable sum rate. To address the non-stationarity introduced by parallel decision-making among the partially observable UAV agents, we propose a teammate modeling based multi-agent reinforcement learning algorithm, named teammate-advisory advantage actor-critic (TAA2C). This algorithm facilitates the exchange of low-dimensional advisory information among the UAV agents to enhance the inter-Uavcollaboration while maintaining low communication overhead. Simulation results demonstrate that, our proposed TAA2C algorithm achieves a better trade-off between communication overhead and cooperation efficiency, compared with the baseline algorithms of Independent A2C (IA2C), Federated A2C (FA2C), and Multi-Agent A2C (MAA2C). As the network scale increases, TAA2C even surpasses the centralized training baseline MAA2C in terms of cumulative sum rate, at only 10% of the communication overhead.

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.

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 Data Updating in Federated Learning (DUFL), 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 Degree of Staleness (DoS) 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.

Federated learning (FL) emerged as a machine learning approach in situations where the privacy of sensitive data needs to be protected. Within the FL framework, clients collaborate to train a shared global model using their individual data by sending the model parameters to a central server, all while keeping their private data localized. Albeit its advantage, FL still faces certain limitations, such as clients may lack the motivation to participate in training, and the heterogeneous data distribution among clients can slow down the model convergence rate and degrade the model accuracy. In the light of the above mentioned considerations, we introduce entropy-based clustered FL (EntroCFL) with incentive mechanism, a two-layer clustered FL (CFL) model to jointly address the incentive mechanism and model training performance issues with heterogeneous clients. In Layer I, the server designs the payments to the clients to minimize its cost, including the training accuracy loss and the payment to clients, based on which the clients determine their training datasizes to maximize their own utilities. In Layer II, we introduce an entropy-based clustering method that is implemented based on the clients’ strategies in Layer I. Unlike conventional CFL methods that rely solely on the cosine similarity between clients’ parameter gradients, EntroCFL introduces a novel clustering discriminant which takes both angle and magnitude of clients’ parameter gradients into consideration. Simulation experiments are conducted to compare EntroCFL with conventional methods, such as FedAvg on the MNIST, EMNIST, and FMNIST datasets. The results validate the superiority of EntroCFL in terms of experimental accuracy, robustness, and economic efficiency.

In federated learning (FL) systems, the central server typically samples a subset of participating clients at each global iteration for model training. To mitigate privacy leakage, clients may insert noise into local parameters before uploading them for global aggregation, leading to FL model performance degradation. This paper aims to design aPrivacy-awareClientSampling framework inFEDerated 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. Through rigorously theoretical convergence analysis, we guarantee the robustness of our proposed FedPCS. Moreover, considering the conventional sampling strategy in privacy-preserving federated learning, we prove that the random sampling approach’s price of anarchy (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 state-of-the-art FL strategies under IID and Non-IID datasets.

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.

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.

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 robustness of our solution against estimation error of clients’ data sizes, and run numerical experiments to validate our results.

Mobile edge computing (MEC) on unmanned aerial vehicles (UAVs) has emerged as a promising method to enhance the computational capabilities of mobile devices (MDs) with limited resources. However, most of the current research on computational offloading algorithms focuses on optimizing the delay and energy of users and lacks attention to the economy of the MEC system. Therefore, in this paper, we propose a three-stage UAV-enabled MEC system model to optimize MDs' cost and the UAV's revenue. To minimize the cost of MDs, we propose a partial computing offloading optimization method based on monetary cost. We investigate a cost-minimization strategy for partial offload that jointly controls task allocation and local central processing unit (CPU) frequency in MDs. In addition, we consider a revenue maximization problem for UAV servers with limited computing resources and propose a heuristic algorithm to determine the optimal service price to solve this problem. Finally, we apply an optimal UAV deployment method to maximize total revenue across all regions. Simulation results demonstrate the effectiveness of our proposed scheme in cost-saving and pricing.

Today many mobile users in various zones are invited to sense and send back real-time useful information 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 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, 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, 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.

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. The Internet service providers (ISPs) directly or commission the third-party UAV firms to provide UAV-provided services (UPS) to improve and make up for the shortage of their current mobile services for additional profit. Yet the UAV has limited energy storage and needs to fly to serve users locally, requiring an optimal energy allocation for balancing both hovering time and service capacity. For profit-maximizing purpose, 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. This paper first introduce a threshold-based assignment policy to show how the UAV decides to serve the users or not under complete information that a user’s service valuation can be observed when he arrives. Following this benchmark, we analyze the UAV’s optimal pricing under incomplete information about the users’ random arrival and private service valuations. It is proved 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. Then, based on the optimal pricing, the energy allocation to hovering time and service capacity in a hotspot is optimized. We show that as the hotspot’s user occurrence rate increases, a shorter hovering time or a larger service capacity should be allocated. Finally, when a 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, by taking the optimal pricing and energy allocation of each hotspot into consideration. With multiple UAVs, however, this result can be reversed with UAVs’ forking deployment to different hotspots, especially when hotspots are more symmetric or the UAV number is large. Perhaps surprisingly, more UAVs may be deployed to the second-best hotspot rather than the first-best one.

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 roll 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.

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.

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 realtime 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 ε-optimal version without recursive computing over time.

Mobile data services are becoming the main driver of a wireless service provider's (WSPs) revenue growth, and two-part tariff data plans (each including a lump-sum fee and a per-unit charge) are usually provided to wireless users. Some users can easily use up their monthly data quota and may pay for costly data over-usage. Motivated by users' diverse usage behavior (more or less than the subscribed data quotas), this paper proposes a new type of user-initiated network for cellular users to trade data plans by leveraging personal hotspots (PHs) with users' smartphones. A user with data surplus can set up a PH and share the cellular data connection to another user with data deficit in the vicinity. Due to users' randomness in data usage, incentive to trade, and user mobility to enter or leave the PH connection range, the analysis on the secondary trading market is challenging. To overcome these issues, we propose a PH-market for users with diverse data usage behaviors and random user mobility to directly trade data as sellers and buyers, by designing a market-clearing price. It is shown that the PH-market greatly saves all users' expected costs when the existence condition of the PH-market is met. Finally, as this PH-market will challenge the WSP's revenue collection (especially the surcharge from users' data over-usage), we analyze the WSP's response to the PH-market and propose two effective countermeasure strategies by either reducing the selling users' data quota in their data plans (the PH-market's supply) or increasing the buying users' data quota (the PH-market's demand). When we have more than one WSP and they are competitive, we show that one WSP can take advantage of the PH-market by indirectly selling more data to the other WSP's users.

In this paper, we investigate the behavior of agents in mean field games where each agent evolves according to a dynamic equation containing the input average and seeks to minimize its long time average (LTA) cost encompassing a population state average (PSA), which is also known as the mean field term. Due to the informational burden resulting from the PSA coupling to the states of all agents, our idea is to find a deterministic function φ to approximate it. It is shown that φ is an approximation of the PSA as the population size N goes to infinity. The resulting decentralized mean field control laws lead the system to achieve mean-consensus asymptotically as time goes to infinity. Furthermore, the optimal controls generate an almost sure asymptotic Nash equilibrium, which implies that the LTA cost of each agent can reach its minimal value as the number of agents increases to infinity. Finally, we consider the socially optimal case where the basic objective is to minimize the social cost as the sum of the individual LTA cost containing the PSA. In this case, it is shown that the decentralized mean field social control strategies are the same as the mean field Nash controls for infinite population systems.