arXiv subjects

Lingjie Duan

Publications and source records attributed to Lingjie Duan.

At least 19 recordsLinked to original sources

arXiv2026

The Forgetting-Retention Dilemma: Certified Unlearning Theory in Continual Learning

Machine unlearning aims to eliminate the influence of specific data from trained models to safeguard privacy. However, this presents a significant challenge in the context of continual learning (CL), where models update sequentially on dynamic datasets. A major limitation is that current certified unlearning algorithms fail to account for the complex, cumulative model evolution inherent to CL framework. In this work, we establish the first theoretical foundation bridging CL and machine unlearning. We formulate the CL's unlearning objective as the minimization of post-unlearning excess risk, which decomposes into CL excess risk and unlearning loss, characterizing the fundamental trade-off between preserving historical knowledge and targeted forgetting. Under mild assumptions, we first establish an upper bound for the CL excess risk in non-convex models. We then adapt two certified unlearning approaches, gradient-based and Hessian-based, to the CL framework. Our analysis reveals that while the gradient-based approach is less effective than the Hessian-based method in minimizing unlearning loss, it offers the distinct advantage of nearly zero storage overhead for enabling unlearning. This insight motivates a hybrid strategy that reduces storage costs while maintaining post-unlearning performance. Experimental results further validate our theoretical findings.

cs.LG↗

arXiv2026

Theory of Continual Learning Against Data Poisoning Attacks

Continual learning (CL), where a model is trained on a sequence of data tasks, is increasingly being adopted across key fields such as large language models and image recognition, yet it remains highly vulnerable to data poisoning that triggers learning divergence or severe excess risk. Despite these threats, a principled theoretical foundation in CL for understanding attack and defense remains lacking. In this paper, we develop a theoretical framework to analyze strategic attacks and defenses in regularization-based CL, a cornerstone of recent CL theory. By framing the adversary-defender interaction as an online zero-sum game, we first establish a fundamental performance limit: no defense succeeds when an adversary poisons a linear proportion of tasks by injecting unbounded noise or pattern shifts in regularization-based CL. We then analyze two possibly defensible scenarios: infrequent attacks and bounded noise per attack. For the former regime, we propose a task-to-task verification mechanism to detect data poisoning and reduce cumulative bias for learning convergence. For the latter regime, we derive a robust defense that minimizes the model's sensitivity to poisoned features, provably accelerating the convergence rate. Extensive experiments on realistic tasks further validate our theoretical results.

cs.LG↗

arXiv2026

When Mobile Crowdsourcing Meets Queueing Systems: Human-in-the-Loop Learning

In service systems, customers now rely on congestion information before deciding which queue or server to join, from restaurants and theme-park attractions to road networks. We study this setting as human-in-the-loop learning (HILL), where customers both consume and generate time-sensitive congestion information through crowdsourcing platforms. Because congestion reports become stale, efficient system operation requires continued exploration of servers whose current states are uncertain. Yet selfish customers avoid such exploration when it reduces their immediate service utility, even though their observations would benefit future customers. We analyze this tension between individual incentives and system-wide learning in queueing systems with endogenous congestion. We first show that myopic server choices can induce an infinite price of anarchy (PoA): decentralized customers may cause arbitrarily large efficiency losses by overexploring servers that are likely congested. In the single-server case, we prove that the lower bound on PoA decreases as buffer size grows, while in the multi-server case the upper bound decreases as the number of servers increases. We further show that existing informational, non-monetary mechanisms for exploration-exploitation with exogenous information fail in our setting, as customers' choices directly reshape the queue states and still lead to infinite PoA. To address this challenge, we design a dynamic side-payment mechanism that periodically charges some customers and rewards others, discouraging excessive exploration while maintaining ex-post budget balance. The mechanism coordinates congestion management and information acquisition across heterogeneous servers, and guarantees PoA below 2. Beyond worst-case analysis, experiments using real datasets demonstrate that the proposed mechanism also achieves strong average-case performance.

cs.GT↗

arXiv2026

Truthful Online Preference Aggregation for LLM Fine-Tuning in Mobile Crowdsourcing

To better serve users' demands in mobile applications (e.g., navigation), mobile crowdsourcing platforms can iteratively align large language model (LLM)-generated content (e.g., AI-generated traffic condition predictions) with human feedback collected from crowdsourcing workers (e.g., mobile users). However, workers may strategically misreport their online preference feedback to maximize their influence or payment. Existing pipelines in mobile crowdsourcing (e.g., EM-based weight estimation) fail to identify the most accurate worker in this online setting, resulting in a linear regret $\mathcal{O}(T)$ over $T$ time slots. In this paper, we study truthful online preference aggregation for LLM fine-tuning in mobile crowdsourcing. We formulate a new dynamic Bayesian game to model the multi-agent online learning process between the platform and strategic mobile workers. We propose a novel online weighted aggregation mechanism that dynamically adjusts each worker's weight in the preference aggregation according to their feedback accuracy. We prove that our mechanism ensures truthful feedback from strategic workers and achieves a sublinear regret $\mathcal{O}(\sqrt{T})$ over $T$ time slots. We further extend our mechanism to a challenging scenario with limited worker feedback per time slot, still guaranteeing a sublinear regret $\mathcal{O}(\sqrt{T})$. Experiments on LLM fine-tuning with real-world datasets further demonstrate significant performance gains of our mechanisms over benchmark schemes.

cs.LG↗

arXiv2026

Learning From Social Interactions: Personalized Pricing and Buyer Manipulation

As the sociological theory of homophily suggests, people tend to interact with those of similar preferences. Motivated by this well-established phenomenon, today's online sellers, such as Amazon,~seek~to learn a new buyer's private preference from his friends' purchase records. Although such learning allows the seller to enable personalized pricing and boost revenue, buyers are also increasingly aware of these practices and may alter their social behaviors accordingly. This paper presents the first study regarding how buyers strategically manipulate their social interaction signals considering their preference correlations, and how a seller can take buyers' strategic social behaviors into consideration when designing the pricing scheme. Starting with the fundamental two-buyer network, we propose and analyze a parsimonious model that uniquely captures the double-layered information asymmetry between the seller and buyers, integrating both individual buyer information and inter-buyer correlation information. Our analysis reveals that only high-preference buyers tend to manipulate their social interactions to evade the seller's personalized pricing, but surprisingly, their payoffs may actually worsen as a result. Moreover, we demonstrate that the seller can considerably benefit from the learning practice, regardless of whether the buyers are aware of this fact or not. Indeed, our analysis reveals that buyers' learning-aware strategic manipulation has only a slight impact on the seller's revenue. In light of the tightening regulatory policies concerning data access, it is advisable for sellers to maintain transparency with buyers regarding their access to buyers' social interaction data for learning purposes. This finding aligns well with current informed-consent industry practices for data sharing.

cs.GT↗

arXiv2026

Governing AI Forgetting: Auditing for Machine Unlearning Compliance

Despite legal mandates for the right to be forgotten, AI operators routinely fail to comply with data deletion requests. While machine unlearning (MU) provides a technical solution to remove personal data's influence from trained models, ensuring compliance remains challenging due to the fundamental gap between MU's technical feasibility and regulatory implementation. In this paper, we introduce the first economic framework for auditing MU compliance, by integrating certified unlearning theory with regulatory enforcement. We first characterize MU's inherent verification uncertainty using a hypothesis-testing interpretation of certified unlearning to derive the auditor's detection capability, and then propose a game-theoretic model to capture the strategic interactions between the auditor and the operator. A key technical challenge arises from MU-specific nonlinearities inherent in the model utility and the detection probability, which create complex strategic couplings that traditional auditing frameworks do not address and that also preclude closed-form solutions. We address this by transforming the complex bivariate nonlinear fixed-point problem into a tractable univariate auxiliary problem, enabling us to decouple the system and establish the equilibrium existence, uniqueness, and structural properties without relying on explicit solutions. Counterintuitively, our analysis reveals that the auditor can optimally reduce the inspection intensity as deletion requests increase, since the operator's weakened unlearning makes non-compliance easier to detect. This is consistent with recent auditing reductions in China despite growing deletion requests. Moreover, we prove that although undisclosed auditing offers informational advantages for the auditor, it paradoxically reduces the regulatory cost-effectiveness relative to disclosed auditing.

cs.LG↗

arXiv2026

ROSS: RObust decentralized Stochastic learning based on Shapley values

In the paradigm of decentralized learning, a group of agents collaborate to learn a global model using a distributed dataset without a central server; nevertheless, it is severely challenged by the heterogeneity of the data distribution across the agents. For example, the data may be distributed non-independently and identically, and even be noised or poisoned. To address these data challenges, we propose ROSS, a novel robust decentralized stochastic learning algorithm based on Shapley values, in this paper. Specifically, in each round, each agent aggregates the cross-gradient information from its neighbors, i.e., the derivatives of its local model with respect to the datasets of its neighbors, to update its local model in a momentum like manner, while we innovate in weighting the derivatives according to their contributions measured by Shapley values. We perform solid theoretical analysis to reveal the linear convergence speedup of our ROSS algorithm. We also verify the efficacy of our algorithm through extensive experiments on public datasets. Our results demonstrate that, in face of the above variety of data challenges, our ROSS algorithm has significant advantages over existing state-of-the-art proposals in terms of both convergence and prediction accuracy.

cs.LG↗

arXiv2025

Personalized Pricing Through Strategic User Profiling in Social Networks

Traditional user profiling techniques rely on browsing history or purchase records to identify users' willingness to pay. This enables sellers to offer personalized prices to profiled users while charging only a uniform price to non-profiled users. However, the emergence of privacy-enhancing technologies has caused users to actively avoid on-site data tracking. Today, major online sellers have turned to public platforms such as online social networks to better track users' profiles from their product-related discussions. This paper presents the first analytical study on how users should best manage their social activities against potential personalized pricing, and how a seller should strategically adjust her pricing scheme to facilitate user profiling in social networks. We formulate a dynamic Bayesian game played between the seller and users under asymmetric information. The key challenge of analyzing this game comes from the double couplings between the seller and the users as well as among the users. Furthermore, the equilibrium analysis needs to ensure consistency between users' revealed information and the seller's belief under random user profiling. We address these challenges by alternately applying backward and forward induction, and successfully characterize the unique perfect Bayesian equilibrium (PBE) in closed form. Our analysis reveals that as the accuracy of profiling technology improves, the seller tends to raise the equilibrium uniform price to motivate users' increased social activities and facilitate user profiling. However, this results in most users being worse off after the informed consent policy is imposed to ensure users' awareness of data access and profiling practices by potential sellers. This finding suggests that recent regulatory evolution towards enhancing users' privacy awareness may have unintended consequences of reducing users' payoffs.

cs.GT↗

arXiv2025

Provable In-Context Learning of Nonlinear Regression with Transformers

The transformer architecture, which processes sequences of input tokens to produce outputs for query tokens, has revolutionized numerous areas of machine learning. A defining feature of transformers is their ability to perform previously unseen tasks using task specific prompts without updating parameters, a phenomenon known as in-context learning (ICL). Recent research has actively explored the training dynamics behind ICL, with much of the focus on relatively simple tasks such as linear regression and binary classification. To advance the theoretical understanding of ICL, this paper investigates more complex nonlinear regression tasks, aiming to uncover how transformers acquire in-context learning capabilities in these settings. We analyze the stage-wise dynamics of attention during training: attention scores between a query token and its target features grow rapidly in the early phase, then gradually converge to one, while attention to irrelevant features decays more slowly and exhibits oscillatory behavior. Our analysis introduces new proof techniques that explicitly characterize how the nature of general non-degenerate $L$-Lipschitz task functions affects attention weights. Specifically, we identify that the Lipschitz constant $L$ of nonlinear function classes as a key factor governing the convergence dynamics of transformers in ICL. Leveraging these insights, for two distinct regimes depending on whether $L$ is below or above a threshold, we derive different time bounds to guarantee near-zero prediction error. Notably, despite the convergence time depending on the underlying task functions, we prove that query tokens consistently attend to prompt tokens with highly relevant features at convergence, demonstrating the ICL capability of transformers for unseen functions.

cs.LG↗

arXiv2025

Human-in-the-loop Learning Through Decentralized Communication Mechanisms

Information sharing platforms like TripAdvisor and Waze involve human agents as both information producers and consumers. All these platforms operate in a centralized way to collect agents' latest observations of new options (e.g., restaurants, hotels, travel routes) and share such information with all in real time. However, after hearing the central platforms' live updates, many human agents are found selfish and unwilling to further explore unknown options for the benefit of others in the long run. To regulate the human-in-the-loop learning (HILL) game against selfish agents' free-riding, this paper proposes a paradigm shift from centralized to decentralized way of operation that forces agents' local explorations through restricting information sharing. When game theory meets distributed learning, we formulate our decentralized communication mechanism's design as a new multi-agent Markov decision process (MA-MDP), and derive its analytical condition to outperform today's centralized operation. As the optimal decentralized communication mechanism in MA-MDP is NP-hard to solve, we present an asymptotically optimal algorithm with linear complexity to determine the mechanism's timing of intermittent information sharing. Then we turn to non-myopic agents who may revert to even over-explore, and adapt our mechanism design to work. Simulation experiments using real-world dataset demonstrate the effectiveness of our decentralized mechanisms for various scenarios.

cs.SI↗

arXiv2025

To Theoretically Understand Transformer-Based In-Context Learning for Optimizing CSMA

The binary exponential backoff scheme is widely used in WiFi 7 and still incurs poor throughput performance under dynamic channel environments. Recent model-based approaches (e.g., non-persistent and $p$-persistent CSMA) simply optimize backoff strategies under a known and fixed node density, still leading to a large throughput loss due to inaccurate node density estimation. This paper is the first to propose LLM transformer-based in-context learning (ICL) theory for optimizing channel access. We design a transformer-based ICL optimizer to pre-collect collision-threshold data examples and a query collision case. They are constructed as a prompt as the input for the transformer to learn the pattern, which then generates a predicted contention window threshold (CWT). To train the transformer for effective ICL, we develop an efficient algorithm and guarantee a near-optimal CWT prediction within limited training steps. As it may be hard to gather perfect data examples for ICL in practice, we further extend to allow erroneous data input in the prompt. We prove that our optimizer maintains minimal prediction and throughput deviations from the optimal values. Experimental results on NS-3 further demonstrate our approach's fast convergence and near-optimal throughput over existing model-based and DRL-based approaches under unknown node densities.

cs.LG↗

arXiv2025

On Incentivizing Social Information Sharing Through Routing Games

Crowdsourcing services, such as Waze, leverage a mass of mobile users to learn massive point-of-interest (PoI) information while traveling and share it as a public good. Given that crowdsourced users mind their travel costs and possess various preferences over the PoI information along different paths, we formulate the problem as a novel non-atomic multi-path routing game with positive network externalities among users in social information sharing. In the absence of any incentive design, our price of anarchy (PoA) analysis shows that users' selfish routing on the path with the lowest cost will limit information diversity and lead to $PoA = 0$ with an arbitrarily large efficiency loss from the social optimum. This motivates us to design effective incentive mechanisms to remedy while upholding desirable properties such as individual rationality, incentive compatibility, and budget balance for practical users. Without requiring a specific user's path preference, we present a non-monetary mechanism called Adaptive Information Restriction (AIR) that reduces non-cooperative users' access to the public good as an indirect penalty, which meets all the desirable properties. By meticulously adapting penalty fractions to the actual user flows along different paths, our AIR achieves non-trivial $PoA = \frac{1}{4}$ with low complexity $O(k\log k+\log m)$, where $k$ and $m$ denote the numbers of involved paths and user types, respectively. If the system can further enable pricing for users, we then propose a new monetary mechanism called Adaptive Side-Payment (ASP), which adaptively charges and rewards users according to their chosen paths, respectively. Our ASP mechanism successively achieves a $PoA = \frac{1}{2}$ with even reduced complexity $O(k\log k)$. Finally, our theoretical findings are well corroborated by our experimental results using a real-world public dataset.

cs.GT↗

arXiv2025

Staying Fresh: Efficient Algorithms for Timely Social Information Distribution

In location-based social networks (LBSNs), users sense urban point-of-interest (PoI) information in the vicinity and share such information with friends in online social networks. Given users' limited social connections and severe lags in disseminating fresh PoI to all, major LBSNs aim to enhance users' social PoI sharing by selecting $k$ out of $m$ users as hotspots and broadcasting their fresh PoI information to the entire user community. This motivates us to study a new combinatorial optimization problem that involves the interplay between an urban sensing network and an online social network. We prove that this problem is NP-hard and also renders existing approximation solutions not viable. Through analyzing the interplay effects between the two networks, we successfully transform the involved PoI-sharing process across two networks to matrix computations for deriving a closed-form objective to hold desirable properties (e.g., submodularity and monotonicity). This finding enables us to develop a polynomial-time algorithm that guarantees a ($1-\frac{m-2}{m}(\frac{k-1}{k})^k$) approximation of the optimum. Furthermore, we allow each selected user to move around and sense more PoI information to share and propose an augmentation-adaptive algorithm with decent performance guarantees. Finally, our theoretical results are corroborated by our simulation findings using both synthetic and real-world datasets.

cs.SI↗

arXiv2025

Truthful mechanisms for linear bandit games with private contexts

The contextual bandit problem, where agents arrive sequentially with personal contexts and the system adapts its arm allocation decisions accordingly, has recently garnered increasing attention for enabling more personalized outcomes. However, in many healthcare and recommendation applications, agents have private profiles and may misreport their contexts to gain from the system. For example, in adaptive clinical trials, where hospitals sequentially recruit volunteers to test multiple new treatments and adjust plans based on volunteers' reported profiles such as symptoms and interim data, participants may misreport severe side effects like allergy and nausea to avoid perceived suboptimal treatments. We are the first to study this issue of private context misreporting in a stochastic contextual bandit game between the system and non-repeated agents. We show that traditional low-regret algorithms, such as UCB family algorithms and Thompson sampling, fail to ensure truthful reporting and can result in linear regret in the worst case, while traditional truthful algorithms like explore-then-commit (ETC) and $ε$-greedy algorithm incur sublinear but high regret. We propose a mechanism that uses a linear program to ensure truthfulness while minimizing deviation from Thompson sampling, yielding an $O(\ln T)$ frequentist regret. Our numerical experiments further demonstrate strong performance in multiple contexts and across other distribution families.

cs.LG↗

arXiv2025

Competitive Multi-armed Bandit Games for Resource Sharing

In modern resource-sharing systems, multiple agents access limited resources with unknown stochastic conditions to perform tasks. When multiple agents access the same resource (arm) simultaneously, they compete for successful usage, leading to contention and reduced rewards. This motivates our study of competitive multi-armed bandit (CMAB) games. In this paper, we study a new N-player K-arm competitive MAB game, where non-myopic players (agents) compete with each other to form diverse private estimations of unknown arms over time. Their possible collisions on same arms and time-varying nature of arm rewards make the policy analysis more involved than existing studies for myopic players. We explicitly analyze the threshold-based structures of social optimum and existing selfish policy, showing that the latter causes prolonged convergence time $Ω(\frac{K}{η^2}\ln({\frac{KN}δ}))$, while socially optimal policy with coordinated communication reduces it to $\mathcal{O}(\frac{K}{Nη^2}\ln{(\frac{K}δ)})$. Based on the comparison, we prove that the competition among selfish players for the best arm can result in an infinite price of anarchy (PoA), indicating an arbitrarily large efficiency loss compared to social optimum. We further prove that no informational (non-monetary) mechanism (including Bayesian persuasion) can reduce the infinite PoA, as the strategic misreporting by non-myopic players undermines such approaches. To address this, we propose a Combined Informational and Side-Payment (CISP) mechanism, which provides socially optimal arm recommendations with proper informational and monetary incentives to players according to their time-varying private beliefs. Our CISP mechanism keeps ex-post budget balanced for social planner and ensures truthful reporting from players, achieving the minimum PoA=1 and same convergence time as social optimum.

cs.GT↗

arXiv2025

Theory of Mixture-of-Experts for Mobile Edge Computing

In mobile edge computing (MEC) networks, mobile users generate diverse machine learning tasks dynamically over time. These tasks are typically offloaded to the nearest available edge server, by considering communication and computational efficiency. However, its operation does not ensure that each server specializes in a specific type of tasks and leads to severe overfitting or catastrophic forgetting of previous tasks. To improve the continual learning (CL) performance of online tasks, we are the first to introduce mixture-of-experts (MoE) theory in MEC networks and save MEC operation from the increasing generalization error over time. Our MoE theory treats each MEC server as an expert and dynamically adapts to changes in server availability by considering data transfer and computation time. Unlike existing MoE models designed for offline tasks, ours is tailored for handling continuous streams of tasks in the MEC environment. We introduce an adaptive gating network in MEC to adaptively identify and route newly arrived tasks of unknown data distributions to available experts, enabling each expert to specialize in a specific type of tasks upon convergence. We derived the minimum number of experts required to match each task with a specialized, available expert. Our MoE approach consistently reduces the overall generalization error over time, unlike the traditional MEC approach. Interestingly, when the number of experts is sufficient to ensure convergence, adding more experts delays the convergence time and worsens the generalization error. Finally, we perform extensive experiments on real datasets in deep neural networks (DNNs) to verify our theoretical results.

cs.LG↗

arXiv2025

Theory on Mixture-of-Experts in Continual Learning

Continual learning (CL) has garnered significant attention because of its ability to adapt to new tasks that arrive over time. Catastrophic forgetting (of old tasks) has been identified as a major issue in CL, as the model adapts to new tasks. The Mixture-of-Experts (MoE) model has recently been shown to effectively mitigate catastrophic forgetting in CL, by employing a gating network to sparsify and distribute diverse tasks among multiple experts. However, there is a lack of theoretical analysis of MoE and its impact on the learning performance in CL. This paper provides the first theoretical results to characterize the impact of MoE in CL via the lens of overparameterized linear regression tasks. We establish the benefit of MoE over a single expert by proving that the MoE model can diversify its experts to specialize in different tasks, while its router learns to select the right expert for each task and balance the loads across all experts. Our study further suggests an intriguing fact that the MoE in CL needs to terminate the update of the gating network after sufficient training rounds to attain system convergence, which is not needed in the existing MoE studies that do not consider the continual task arrival. Furthermore, we provide explicit expressions for the expected forgetting and overall generalization error to characterize the benefit of MoE in the learning performance in CL. Interestingly, adding more experts requires additional rounds before convergence, which may not enhance the learning performance. Finally, we conduct experiments on both synthetic and real datasets to extend these insights from linear models to deep neural networks (DNNs), which also shed light on the practical algorithm design for MoE in CL.

cs.LG↗

arXiv2025

Mechanism Design for Blockchain Order Books against Selfish Miners

In blockchain-based order book systems, buyers and sellers trade assets, while it is miners to match them and include their transactions in the blockchain. It is found that many miners behave selfishly and myopically, prioritizing transactions with high fees and ignoring many desirable matches that could enhance social welfare. Existing blockchain mechanisms fail to address this issue by overlooking miners' selfish behaviors. To our best knowledge, this work presents the first analytical study to quantify and understand buyer and seller transaction fee choices and selfish miners' transaction matching strategies, proving an infinitely large price of anarchy (PoA) for social welfare loss. To mitigate this, we propose an adjustable block size mechanism that is easy to implement without altering the existing decentralized protocols and still allows buyers and sellers to freely decide transaction fees and miners to selfishly match. The analysis is challenging, as pure strategy Nash equilibria do not always exist, requiring the analysis of many buyers' or sellers' interactive mixed-strategy distributions. Moreover, the system designer may even lack information about each buyer's or seller's bid/ask prices and trading quantities. Nevertheless, our mechanism achieves a well-bounded PoA, and under the homogeneous-quantity trading for non-fungible tokens (NFT), it attains a PoA of 1 with no social welfare loss. We implement our mechanism on a local instance of Ethereum to demonstrate the feasibility of our approach. Experiments based on the realistic dataset demonstrate that our mechanism achieves social optimum for homogeneous-quantity trading like NFT. It can enhance social welfare up to 3.7 times compared to the existing order book benchmarks for heterogeneous-quantity trading of Bitcoin tokens. It exhibits robustness against random variations in buyers and sellers.

cs.GT↗