arXiv subjects

Sujit Gujar

Publications and source records attributed to Sujit Gujar.

At least 19 recordsLinked to original sources

arXiv2026

Fair Stable Matching: A Nash Social Welfare Approach

While traditional stable matching algorithms, such as the Gale-Shapley algorithm, prioritize stability, they may fall short of achieving equitable outcomes among participants. We study the role of \emph{Nash social welfare} (NSW) as a fairness objective in the classic \emph{stable marriage problem}. We develop \texttt{SNSW-Alg} that finds a stable matching that maximizes Nash social welfare under rank-induced utilities in $\tilde{\mathcal{O}}(n^4)$ time, where $n$ is the number of men or women. We demonstrate that \texttt{SNSW-Alg} balances equity while preserving stability. We empirically evaluate our methods across diverse preference distributions, demonstrating significant gains in fairness without substantial losses in other key measures such as regret, egalitarian criterion, and sex equality. Our findings suggest that the stable matching produced by \texttt{SNSW-Alg} is statistically Pareto-undominated by stable matchings based on other fairness measures - regret, egalitarian, and sex equality. This study offers compelling insights for designing fair-stable matching.

cs.GT↗

arXiv2026

Resourced Authority A Mechanism-Design Model for Participatory Governance of Deployed AI Agents

We give a formal mechanism design model for the continuous participatory governance of a deployed AI agent. The mechanism is built on the principle that governance should control an AI agent through resource allocation so as to make authorization self enforcing via compute budgets. The mechanism seeks to establish the Safe AI paradigm that compute is an effective governance lever. We situate our work as a compliance or commons overlay on a deployer. One governance period is an extensive form game in which verified human stakeholders arrive sequentially and contribute, on a provision or a rejection market, in a governance currency that is deliberately distinct from the agents compute. A funding aggregator turns raw contributions into breadth weighted effective supports - a two threshold gate with hysteresis converts net support into a binary authorization that, through a coupling map bounded by an exogenously certified safety ceiling, releases a metered compute budget - realized in hardware as a signed compute license so that the decision is self-enforcing. We characterize the class of agents the mechanism can govern and isolate manipulation of the governing electorate by the governed agent as the central open problem. We also introduce several challenges addressing manipulation of governing electorate by the governed agents.

cs.GT↗

arXiv2026

Fairness Auditing: Lower Bounds on Company Manipulation

Fairness audits are increasingly mandated in high-stakes applications such as hiring, lending, and automated decision-making. Recent work has established fundamental impossibility results for black-box fairness auditing, showing that sufficiently expressive models can evade any auditing strategy. We complement these results by quantifying the extent of unavoidable post-audit manipulation under finite audit resources. We formulate fairness auditing as a min-max optimization between a computationally unbounded company and a budget-constrained auditor. We study two auditing regimes: (i) a budgeted auditor that certifies fairness using a fixed-size audit set, and (ii) a budgeted {\alpha}-tolerant auditor that additionally requires the audit set to estimate the fairness of the certified model within an {\alpha} approximation. For both settings, we derive explicit lower bounds on the worst-case post-audit demographic parity deviation as functions of the audit budget, group imbalance, and fairness tolerance. Finally, we empirically illustrate these theoretical limits using simple audit-set construction heuristics with linear and neural network classifiers. Our results demonstrate that increasing audit resources reduces, but does not eliminate, the scope for post-audit manipulation, highlighting fundamental limitations of finite-budget fairness certification.

cs.LG↗

arXiv2026

Multi-Sender Bayesian Persuasion with Imperfect Information

We study a multi-sender Bayesian persuasion problem with one receiver and several strategic senders. The underlying ground state has multiple components, each privately observed by a different sender, while the receiver holds a common prior over the joint state space. Senders simultaneously choose signaling policies, and the receiver takes an action based on the posterior induced by the signals; each is sampled independently from the sender's signaling policy. We analyze the game induced by the receiver's straightforward policy, which selects a receiver-optimal action at every posterior. In particular, we characterize the senders' best responses under the straightforward policy and identify conditions on the prior that induce a fully informative equilibrium; i.e., truthfully reporting the ground truth is an equilibrium strategy for every sender. These conditions capture cases in which senders' incentives are sufficiently aligned to enable full revelation without additional commitment from the receiver. The important contribution of this paper is to analyze games induced by a more general (possibly randomized) class of action policies that the receiver commits to before senders choose their signaling strategies. We show that this commitment power fundamentally changes the problem. In particular, we show that for any prior over the joint state space, the receiver can construct action policies that maximize her payoff while ensuring a fully informative equilibrium.

cs.GT↗

arXiv2026

Linear Strategic Classification with Endogenous Improvements

Strategic classification studies settings in which agents respond to a deployed classifier by modifying observable features at a cost. Classical models typically treat such responses as cosmetic: features may change, but true labels remain fixed. We study an improvement-aware variant in which strategic responses can induce genuine changes in outcome-relevant features. Agents choose post-deployment feature vectors strategically, and labels are then generated according to a stable conditional outcome law that preserves the relationship between features and outcomes. We formalize this problem for linear classifiers under a single-index qualification model and linear-decomposable costs. We show that the strategic-optimal classifier is obtained by a parallel shift of the Bayes-optimal decision boundary, and that it provides a better surrogate for the improvement-aware objective than the Bayes classifier. Since improvement-aware learning requires post-deployment labels, which are typically unavailable before deployment, we provide PAC-style guar- antees under an oracle model, propose a practical plug-in algorithm, establish its generalization bound, and evaluate it on synthetic and real-world datasets.

cs.LG↗

arXiv2026

Credibility Trilemma in Polymatroidal Service Markets

Mechanism-mediated service markets with polymatroidal feasibility admit efficient, dominant-strategy incentive-compatible (DSIC) allocation, but these guarantees implicitly assume truthful execution by the marketplace operator. Modelling the operator as a strategic player, we establish a credibility trilemma: for single-parameter agents on a non-modular polymatroid carrying a competitive prior profile, no static sealed-bid mechanism is simultaneously revenue-optimal, DSIC for agents, and credible for the operator. We introduce the Cost of Non-Credibility (CoNC) as a price-of-anarchy-style welfare-loss measure and prove a per-pair lower bound scaling with the polymatroid's non-modularity gap, turning the trilemma into a quantitative diagnostic. Three structurally distinct resolutions follow: public broadcast or deferred-revelation commitment, administrative domain separation under settlement separation and four side conditions, and integrator competition orthogonal to mechanism execution under disjoint actors. An instance-level grounding over the edge-pricing market of Amin et al. shows that the economic content carries over to a refereed external setting. Marketplace neutrality is thus a first-order design constraint on polymatroidal service markets rather than an implementation detail: where the operator is a strategic player, credibility trades off against revenue optimality and agent incentive compatibility along structurally characterised lines.

cs.GT↗

arXiv2026

Truthful Reverse Auctions for Adaptive Selection via Contextual Multi-Armed Bandits

We study the problem of selecting large language models (LLMs) for user queries in settings where multiple LLM providers submit the cost of solving a query. From the users' perspective, choosing an optimal model is a sequential, query-dependent decision problem: high-capacity models offer more reliable outputs but are costlier, while lightweight models are faster and cheaper. We formalize this interaction as a reverse auction design problem with contextual online learning, where the user adaptively discovers which model performs best while eliciting costs from competing LLM providers. Existing multi-armed bandit (MAB) mechanisms focus on forward auctions and social welfare, leaving open the challenges of reverse auctions, provider-optimal outcomes, and contextual adaptation. We address these gaps by designing a resampling-based procedure that generalizes truthful forward MAB mechanisms to reverse auctions and prove that any monotone allocation rule with this procedure is truthful. Using this, we propose a contextual MAB algorithm that learns query-dependent model quality with sublinear regret. Our framework unifies mechanism design and adaptive learning, enabling efficient, truthful, and query-aware LLM selection.

cs.GT↗

arXiv2026

Let Leaders Play Games: Improving Timing in Leader-based Consensus

Propagation latency is inherent to any distributed network, including blockchains. Typically, blockchain protocols provide a timing buffer for block propagation across the network. In leader-based blockchains, the leader -- block proposer -- is known in advance for each slot. A fast (or low-latency) proposer may delay the block proposal in anticipation of more rewards from the transactions that would otherwise be included in the subsequent block. Deploying such a strategy by manipulating the timing is known as timing games. It increases the risk of missed blocks due to reduced time for other nodes to vote on the block, affecting the overall efficiency of the blockchain. Moreover, proposers who play timing games essentially appropriate MEV (additional rewards over transaction fees and the block reward) that would otherwise accrue to the next block, making it unfair to subsequent block proposers. We propose a double-block proposal mechanism, 2-Prop, to curtail timing games. 2-Prop selects two proposers per slot to propose blocks and confirms one of them. We design a reward-sharing policy for proposers based on how quickly their blocks propagate to avoid strategic deviations. In the induced game, which we call the Latency Game, we show that it is a Nash Equilibrium for the proposers to propose the block without delay under homogeneous network settings. Under heterogeneous network settings, we study many configurations, and our analysis shows that a faster proposer would prefer not to delay unless the other proposer is extremely slow. Thus, we show the efficacy of 2-Prop in mitigating the effect of timing games.

cs.GT↗

arXiv2025

Shapley Value-based Approach for Redistributing Revenue of Matchmaking of Private Transactions in Blockchains

In the context of blockchain, MEV refers to the maximum value that can be extracted from block production through the inclusion, exclusion, or reordering of transactions. Searchers often participate in order flow auctions (OFAs) to obtain exclusive rights to private transactions, available through entities called matchmakers, also known as order flow providers (OFPs). Most often, redistributing the revenue generated through such auctions among transaction creators is desirable. In this work, we formally introduce the matchmaking problem in MEV, its desirable properties, and associated challenges. Using cooperative game theory, we formalize the notion of fair revenue redistribution in matchmaking and present its potential possibilities and impossibilities. Precisely, we define a characteristic form game, referred to as RST-Game, for the transaction creators. We propose to redistribute the revenue using the Shapley value of RST-Game. We show that the corresponding problem could be SUBEXP (i.e. $2^{o(n)}$, where $n$ is the number of transactions); therefore, approximating the Shapley value is necessary. Further, we propose a randomized algorithm for computing the Shapley value in RST-Game and empirically verify its efficacy.

cs.GT↗

arXiv2025

FLIGHT: Facility Location Integrating Generalized, Holistic Theory of Welfare

The Facility Location Problem (FLP) is a well-studied optimization problem with applications in many real-world scenarios. Past literature has explored the solutions from different perspectives to tackle FLPs. These include investigating FLPs under objective functions such as utilitarian, egalitarian, Nash welfare, etc. Also, there is no treatment for asymmetric welfare functions around the facility. We propose a unified framework, FLIGHT, to accommodate a broad class of welfare notions. The framework undergoes rigorous theoretical analysis, and we prove some structural properties of the solution to FLP. Additionally, we provide approximation bounds, which provide insight into an interesting fact: as the number of agents arbitrarily increases, the choice of welfare notion is irrelevant. Furthermore, the paper also includes results around concentration bounds under certain distributional assumptions over the preferred locations of agents.

cs.GT↗

arXiv2025

Optimal Strategies for Federated Learning Maintaining Client Privacy

Federated Learning (FL) emerged as a learning method to enable the server to train models over data distributed among various clients. These clients are protective about their data being leaked to the server, any other client, or an external adversary, and hence, locally train the model and share it with the server rather than sharing the data. The introduction of sophisticated inferencing attacks enabled the leakage of information about data through access to model parameters. To tackle this challenge, privacy-preserving federated learning aims to achieve differential privacy through learning algorithms like DP-SGD. However, such methods involve adding noise to the model, data, or gradients, reducing the model's performance. This work provides a theoretical analysis of the tradeoff between model performance and communication complexity of the FL system. We formally prove that training for one local epoch per global round of training gives optimal performance while preserving the same privacy budget. We also investigate the change of utility (tied to privacy) of FL models with a change in the number of clients and observe that when clients are training using DP-SGD and argue that for the same privacy budget, the utility improved with increased clients. We validate our findings through experiments on real-world datasets. The results from this paper aim to improve the performance of privacy-preserving federated learning systems.

cs.LG↗

arXiv2024

FROC: Building Fair ROC from a Trained Classifier

This paper considers the problem of fair probabilistic binary classification with binary protected groups. The classifier assigns scores, and a practitioner predicts labels using a certain cut-off threshold based on the desired trade-off between false positives vs. false negatives. It derives these thresholds from the ROC of the classifier. The resultant classifier may be unfair to one of the two protected groups in the dataset. It is desirable that no matter what threshold the practitioner uses, the classifier should be fair to both the protected groups; that is, the $\mathcal{L}_p$ norm between FPRs and TPRs of both the protected groups should be at most $\varepsilon$. We call such fairness on ROCs of both the protected attributes $\varepsilon_p$-Equalized ROC. Given a classifier not satisfying $\varepsilon_1$-Equalized ROC, we aim to design a post-processing method to transform the given (potentially unfair) classifier's output (score) to a suitable randomized yet fair classifier. That is, the resultant classifier must satisfy $\varepsilon_1$-Equalized ROC. First, we introduce a threshold query model on the ROC curves for each protected group. The resulting classifier is bound to face a reduction in AUC. With the proposed query model, we provide a rigorous theoretical analysis of the minimal AUC loss to achieve $\varepsilon_1$-Equalized ROC. To achieve this, we design a linear time algorithm, namely \texttt{FROC}, to transform a given classifier's output to a probabilistic classifier that satisfies $\varepsilon_1$-Equalized ROC. We prove that under certain theoretical conditions, \texttt{FROC}\ achieves the theoretical optimal guarantees. We also study the performance of our \texttt{FROC}\ on multiple real-world datasets with many trained classifiers.

cs.LG↗

arXiv2024

MEV Ecosystem Evolution From Ethereum 1.0

Smart contracts led to the emergence of the decentralized finance (DeFi) marketplace within blockchain ecosystems, where diverse participants engage in financial activities. In traditional finance, there are possibilities to create values, e.g., arbitrage offers to create value from market inefficiencies or front-running offers to extract value for the participants having privileged roles. Such opportunities are readily available -- searching programmatically in DeFi. It is commonly known as Maximal Extractable Value (MEV) in the literature. In this survey, first, we show how lucrative such opportunities can be. Next, we discuss how protocol-following participants trying to capture such opportunities threaten to sabotage blockchain's performance and the core tenets of decentralization, transparency, and trustlessness that blockchains are based on. Then, we explain different attempts by the community in the past to address these issues and the problems introduced by these solutions. Finally, we review the current state of research trying to restore trustlessness and decentralization to provide all DeFi participants with a fair marketplace.

cs.CR↗

arXiv2024

Towards Rational Consensus in Honest Majority

Distributed consensus protocols reach agreement among $n$ players in the presence of $f$ adversaries; different protocols support different values of $f$. Existing works study this problem for different adversary types (captured by threat models). There are three primary threat models: (i) Crash fault tolerance (CFT), (ii) Byzantine fault tolerance (BFT), and (iii) Rational fault tolerance (RFT), each more general than the previous. Agreement in repeated rounds on both (1) the proposed value in each round and (2) the ordering among agreed-upon values across multiple rounds is called Atomic BroadCast (ABC). ABC is more generalized than consensus and is employed in blockchains. This work studies ABC under the RFT threat model. We consider $t$ byzantine and $k$ rational adversaries among $n$ players. We also study different types of rational players based on their utility towards (1) liveness attack, (2) censorship or (3) disagreement (forking attack). We study the problem of ABC under this general threat model in partially-synchronous networks. We show (1) ABC is impossible for $n/3< (t+k) <n/2$ if rational players prefer liveness or censorship attacks and (2) the consensus protocol proposed by Ranchal-Pedrosa and Gramoli cannot be generalized to solve ABC due to insecure Nash equilibrium (resulting in disagreement). For ABC in partially synchronous network settings, we propose a novel protocol \textsf{pRFT}(practical Rational Fault Tolerance). We show \textsf{pRFT} achieves ABC if (a) rational players prefer only disagreement attacks and (b) $t < \frac{n}{4}$ and $(t + k) < \frac{n}{2}$. In \textsf{pRFT}, we incorporate accountability (capturing deviating players) within the protocol by leveraging honest players. We also show that the message complexity of \textsf{pRFT} is at par with the best consensus protocols that guarantee accountability.

cs.GT↗

arXiv2024

Towards Fairness in Provably Communication-Efficient Federated Recommender Systems

To reduce the communication overhead caused by parallel training of multiple clients, various federated learning (FL) techniques use random client sampling. Nonetheless, ensuring the efficacy of random sampling and determining the optimal number of clients to sample in federated recommender systems (FRSs) remains challenging due to the isolated nature of each user as a separate client. This challenge is exacerbated in models where public and private features can be separated, and FL allows communication of only public features (item gradients). In this study, we establish sample complexity bounds that dictate the ideal number of clients required for improved communication efficiency and retained accuracy in such models. In line with our theoretical findings, we empirically demonstrate that RS-FairFRS reduces communication cost (~47%). Second, we demonstrate the presence of class imbalance among clients that raises a substantial equity concern for FRSs. Unlike centralized machine learning, clients in FRS can not share raw data, including sensitive attributes. For this, we introduce RS-FairFRS, first fairness under unawareness FRS built upon random sampling based FRS. While random sampling improves communication efficiency, we propose a novel two-phase dual-fair update technique to achieve fairness without revealing protected attributes of active clients participating in training. Our results on real-world datasets and different sensitive features illustrate a significant reduction in demographic bias (~approx40\%), offering a promising path to achieving fairness and communication efficiency in FRSs without compromising the overall accuracy of FRS.

cs.IR↗

arXiv2024

Centralization in Proof-of-Stake Blockchains: A Game-Theoretic Analysis of Bootstrapping Protocols

Proof-of-stake (PoS) has emerged as a natural alternative to the resource-intensive Proof-of-Work (PoW) blockchain, as was recently seen with the Ethereum Merge. PoS-based blockchains require an initial stake distribution among the participants. Typically, this initial stake distribution is called bootstrapping. This paper argues that existing bootstrapping protocols are prone to centralization. To address centralization due to bootstrapping, we propose a novel game $\Gamma_\textsf{bootstrap}$. Next, we define three conditions: (i) Individual Rationality (IR), (ii) Incentive Compatibility (IC), and (iii) $(\tau,\delta,\epsilon)-$ Decentralization that an \emph{ideal} bootstrapping protocol must satisfy. $(\tau,\delta,\epsilon)$ are certain parameters to quantify decentralization. Towards this, we propose a novel centralization metric, C-NORM, to measure centralization in a PoS System. We define a centralization game -- $\Gamma_\textsf{cent}$, to analyze the efficacy of centralization metrics. We show that C-NORM effectively captures centralization in the presence of strategic players capable of launching Sybil attacks. With C-NORM, we analyze popular bootstrapping protocols such as Airdrop and Proof-of-Burn (PoB) and prove that they do not satisfy IC and IR, respectively. Motivated by the Ethereum Merge, we study W2SB (a PoW-based bootstrapping protocol) and prove it is ideal. In addition, we conduct synthetic simulations to empirically validate that W2SB bootstrapped PoS is decentralized.

cs.GT↗

arXiv2024

Fairness of Exposure in Online Restless Multi-armed Bandits

Restless multi-armed bandits (RMABs) generalize the multi-armed bandits where each arm exhibits Markovian behavior and transitions according to their transition dynamics. Solutions to RMAB exist for both offline and online cases. However, they do not consider the distribution of pulls among the arms. Studies have shown that optimal policies lead to unfairness, where some arms are not exposed enough. Existing works in fairness in RMABs focus heavily on the offline case, which diminishes their application in real-world scenarios where the environment is largely unknown. In the online scenario, we propose the first fair RMAB framework, where each arm receives pulls in proportion to its merit. We define the merit of an arm as a function of its stationary reward distribution. We prove that our algorithm achieves sublinear fairness regret in the single pull case $O(\sqrt{T\ln T})$, with $T$ being the total number of episodes. Empirically, we show that our algorithm performs well in the multi-pull scenario as well.

cs.LG↗

arXiv2024

Simultaneously Achieving Group Exposure Fairness and Within-Group Meritocracy in Stochastic Bandits

Existing approaches to fairness in stochastic multi-armed bandits (MAB) primarily focus on exposure guarantee to individual arms. When arms are naturally grouped by certain attribute(s), we propose Bi-Level Fairness, which considers two levels of fairness. At the first level, Bi-Level Fairness guarantees a certain minimum exposure to each group. To address the unbalanced allocation of pulls to individual arms within a group, we consider meritocratic fairness at the second level, which ensures that each arm is pulled according to its merit within the group. Our work shows that we can adapt a UCB-based algorithm to achieve a Bi-Level Fairness by providing (i) anytime Group Exposure Fairness guarantees and (ii) ensuring individual-level Meritocratic Fairness within each group. We first show that one can decompose regret bounds into two components: (a) regret due to anytime group exposure fairness and (b) regret due to meritocratic fairness within each group. Our proposed algorithm BF-UCB balances these two regrets optimally to achieve the upper bound of $O(\sqrt{T})$ on regret; $T$ being the stopping time. With the help of simulated experiments, we further show that BF-UCB achieves sub-linear regret; provides better group and individual exposure guarantees compared to existing algorithms; and does not result in a significant drop in reward with respect to UCB algorithm, which does not impose any fairness constraint.

cs.LG↗