arXiv subjects

Sujit Gujar

Publications and source records attributed to Sujit Gujar.

At least 55 records · Page 3Linked to original sources

arXiv2022

F3: Fair and Federated Face Attribute Classification with Heterogeneous Data

Fairness across different demographic groups is an essential criterion for face-related tasks, Face Attribute Classification (FAC) being a prominent example. Apart from this trend, Federated Learning (FL) is increasingly gaining traction as a scalable paradigm for distributed training. Existing FL approaches require data homogeneity to ensure fairness. However, this assumption is too restrictive in real-world settings. We propose F3, a novel FL framework for fair FAC under data heterogeneity. F3 adopts multiple heuristics to improve fairness across different demographic groups without requiring data homogeneity assumption. We demonstrate the efficacy of F3 by reporting empirically observed fairness measures and accuracy guarantees on popular face datasets. Our results suggest that F3 strikes a practical balance between accuracy and fairness for FAC.

cs.LG↗

arXiv2022

Tiramisu: Layering Consensus Protocols for Scalable and Secure Blockchains

Cryptocurrencies are poised to revolutionize the modern economy by democratizing commerce. These currencies operate on top of blockchain-based distributed ledgers. Existing permissionless blockchain-based protocols offer unparalleled benefits like decentralization, anonymity, and transparency. However, these protocols suffer in performance which hinders their widespread adoption. In particular, high time-to-finality and low transaction rates keep them from replacing centralized payment systems such as the Visa network. Permissioned blockchain protocols offer attractive performance guarantees, but they are not considered suitable for deploying decentralized cryptocurrencies due to their centralized nature. Researchers have developed several multi-layered blockchain protocols that combine both permissioned and permissionless blockchain protocols to achieve high performance along with decentralization. The key idea with existing layered blockchain protocols in literature is to divide blockchain operations into two layers and use different types of blockchain protocols to manage each layer. However, many such works come with the assumptions of honest majority which may not accurately reflect the real world where the participants may be self-interested or rational. These assumptions may render the protocols susceptible to security threats in the real world, as highlighted by the literature focused on exploring game-theoretic attacks on these protocols. We generalize the "layered" approach taken by existing protocols in the literature and present a framework to analyze the system in the BAR Model and provide a generalized game-theoretic analysis of such protocols. Using our analysis, we identify the critical system parameters required for a distributed ledger's secure operation in a more realistic setting.

cs.CR↗

arXiv2022

Fair Allocation with Special Externalities

Most of the existing algorithms for fair division do not consider externalities. Under externalities, the utility an agent obtains depends not only on its allocation but also on the allocation of other agents. An agent has a positive (negative) value for the assigned goods (chores). This work focuses on a special case of externality, i.e., an agent receives positive or negative value for unassigned items independent of which other agent gets it. We show that it is possible to adapt existing algorithms using a transformation to ensure certain fairness and efficiency notions in this setting. Despite the positive results, fairness notions like proportionality need to be re-defined. Further, we prove that maximin share (MMS) may not have any multiplicative approximation in this setting. Studying this domain is a stepping stone towards full externalities where ensuring fairness is much more challenging.

cs.GT↗

arXiv2022

Individual Fairness in Feature-Based Pricing for Monopoly Markets

We study fairness in the context of feature-based price discrimination in monopoly markets. We propose a new notion of individual fairness, namely, α-fairness, which guarantees that individuals with similar features face similar prices. First, we study discrete valuation space and give an analytical solution for optimal fair feature-based pricing. We show that the cost of fair pricing is defined as the ratio of expected revenue in an optimal feature-based pricing to the expected revenue in an optimal fair feature-based pricing (CoF) can be arbitrarily large in general. When the revenue function is continuous and concave with respect to the prices, we show that one can achieve CoF strictly less than 2, irrespective of the model parameters. Finally, we provide an algorithm to compute fair feature-based pricing strategy that achieves this CoF.

cs.GT↗

arXiv2022

Multi-unit Double Auctions: Equilibrium Analysis and Bidding Strategy using DDPG in Smart-grids

Periodic double auctions (PDA) have applications in many areas such as in e-commerce, intra-day equity markets, and day-ahead energy markets in smart-grids. While the trades accomplished using PDAs are worth trillions of dollars, finding a reliable bidding strategy in such auctions is still a challenge as it requires the consideration of future auctions. A participating buyer in a PDA has to design its bidding strategy by planning for current and future auctions. Many equilibrium-based bidding strategies proposed are complex to use in real-time. In the current exposition, we propose a scale-based bidding strategy for buyers participating in PDA. We first present an equilibrium analysis for single-buyer single-seller multi-unit single-shot k-Double auctions. Specifically, we analyze the situation when a seller and a buyer trade two identical units of quantity in a double auction where both the buyer and the seller deploy a simple, scale-based bidding strategy. The equilibrium analysis becomes intractable as the number of participants increases. To be useful in more complex settings such as wholesale markets in smart-grids, we model equilibrium bidding strategy as a learning problem. We develop a deep deterministic policy gradient (DDPG) based learning strategy, DDPGBBS, for a participating agent in PDAs to suggest an action at any auction instance. DDPGBBS, which empirically follows the obtained theoretical equilibrium, is easily extendable when the number of buyers/sellers increases. We take Power Trading Agent Competition's (PowerTAC) wholesale market PDA as a testbed to evaluate our novel bidding strategy. We benchmark our DDPG based strategy against several baselines and state-of-the-art bidding strategies of the PowerTAC wholesale market PDA and demonstrate the efficacy of DDPGBBS against several benchmarked strategies.

cs.GT↗

arXiv2022

Budgeted Combinatorial Multi-Armed Bandits

We consider a budgeted combinatorial multi-armed bandit setting where, in every round, the algorithm selects a super-arm consisting of one or more arms. The goal is to minimize the total expected regret after all rounds within a limited budget. Existing techniques in this literature either fix the budget per round or fix the number of arms pulled in each round. Our setting is more general where based on the remaining budget and remaining number of rounds, the algorithm can decide how many arms to be pulled in each round. First, we propose CBwK-Greedy-UCB algorithm, which uses a greedy technique, CBwK-Greedy, to allocate the arms to the rounds. Next, we propose a reduction of this problem to Bandits with Knapsacks (BwK) with a single pull. With this reduction, we propose CBwK-LPUCB that uses PrimalDualBwK ingeniously. We rigorously prove regret bounds for CBwK-LP-UCB. We experimentally compare the two algorithms and observe that CBwK-Greedy-UCB performs incrementally better than CBwK-LP-UCB. We also show that for very high budgets, the regret goes to zero.

cs.LG↗

arXiv2021

REFORM: Reputation Based Fair and Temporal Reward Framework for Crowdsourcing

Crowdsourcing is an effective method to collect data by employing distributed human population. Researchers introduce appropriate reward mechanisms to incentivize agents to report accurately. In particular, this paper focuses on Peer-Based Mechanisms (PBMs). We observe that with PBMs, crowdsourcing systems may not be fair, i.e., agents may not receive the deserved rewards despite investing efforts and reporting truthfully. Unfair rewards for the agents may discourage participation. This paper aims to build a general framework that assures fairness for PBMs in temporal settings, i.e., settings that prefer early reports. Towards this, we introduce two general notions of fairness for PBMs, namely gamma-fairness and qualitative fairness. To satisfy these notions, our framework provides trustworthy agents with additional chances of pairing. We introduce Temporal Reputation Model (TERM) to quantify agents' trustworthiness across tasks. With TERM as the key constituent, we present our iterative framework, REFORM, that can adopt the reward scheme of any existing PBM. We demonstrate REFORM's significance by deploying the framework with RPTSC's reward scheme. Specifically, we prove that REFORM with RPTSC considerably improves fairness; while incentivizing truthful and early reports. We conduct synthetic simulations and show that our framework provides improved fairness over RPTSC.

cs.GT↗

arXiv2021

Mechanism Design without Money for Fair Allocations

Fairness is well studied in the context of resource allocation. Researchers have proposed various fairness notions like envy-freeness (EF), and its relaxations, proportionality and max-min share (MMS). There is vast literature on the existential and computational aspects of such notions. While computing fair allocations, any algorithm assumes agents' truthful reporting of their valuations towards the resources. Whereas in real-world web-based applications for fair division, the agents involved are strategic and may manipulate for individual utility gain. In this paper, we study strategy-proof mechanisms without monetary transfer, which satisfies the various fairness criteria. We know that for additive valuations, designing truthful mechanisms for EF, MMS and proportionality is impossible. Here we show that there cannot be a truthful mechanism for EFX and the existing algorithms for EF1 are manipulable. We then study the particular case of single-minded agents. For this case, we provide a Serial Dictatorship Mechanism that is strategy-proof and satisfies all the fairness criteria except EF.

cs.GT↗

arXiv2021

EEF1-NN: Efficient and EF1 allocations through Neural Networks

Neural networks have shown state-of-the-art performance in designing auctions, where the network learns the optimal allocations and payment rule to ensure desirable properties. Motivated by the same, we focus on learning fair division of resources, with no payments involved. Our goal is to allocate the items, goods and/or chores efficiently among the fair allocations. By fair, we mean an allocation that is Envy-free (EF). However, such an allocation may not always exist for indivisible resources. Therefore, we consider the relaxed notion, Envy-freeness up to one item (EF1) that is guaranteed to exist. However, it is not enough to guarantee EF1 since the allocation of empty bundles is also EF1. Hence, we add the further constraint of efficiency, maximum utilitarian social welfare (USW). In general finding, USW allocations among EF1 is an NP-Hard problem even when valuations are additive. In this work, we design a network for this task which we refer to as EEF1-NN. We propose an UNet inspired architecture, Lagrangian loss function, and training procedure to obtain desired results. We show that EEF1-NN finds allocation close to optimal USW allocation and ensures EF1 with a high probability for different distributions over input valuations. Compared to existing approaches EEF1-NN empirically guarantees higher USW. Moreover, EEF1-NN is scalable and determines the allocations much faster than solving it as a constrained optimization problem.

cs.GT↗

arXiv2021

Designing Refund Bonus Schemes for Provision Point Mechanism in Civic Crowdfunding

Civic crowdfunding (CC) is a popular medium for raising funds for civic projects from interested agents. With Blockchains gaining traction, we can implement CC in a reliable, transparent, and secure manner with smart contracts (SCs). The fundamental challenge in CC is free-riding. PPR, the proposal by Zubrickas [23] of giving refund bonus to the contributors, in the case of the project not getting provisioned, has attractive properties. However, as observed by Chandra et al. [7], PPR faces a challenge wherein the agents defer their contribution until the deadline. We define this delaying of contributions as a race condition. To address this, their proposal, PPS, considers the temporal aspects of a contribution. However, PPS is computationally complex, expensive to implement as an SC, and it being sophisticated, it is difficult to explain to a layperson. In this work, our goal is to identify all essential properties a refund bonus scheme must satisfy in order to curb free-riding while avoiding the race condition. We prove Contribution Monotonicity and Time Monotonicity are sufficient conditions for this. We propose three elegant refund bonus schemes satisfying these two conditions leading to three novel mechanisms for CC - PPRG, PPRE, and PPRP. We show that PPRG is the most cost-effective mechanism when deployed as an SC. We show that under certain modest assumptions on valuations of the agents, in PPRG, the project is funded at equilibrium.

cs.GT↗

arXiv2021

Federated Learning Meets Fairness and Differential Privacy

Deep learning's unprecedented success raises several ethical concerns ranging from biased predictions to data privacy. Researchers tackle these issues by introducing fairness metrics, or federated learning, or differential privacy. A first, this work presents an ethical federated learning model, incorporating all three measures simultaneously. Experiments on the Adult, Bank and Dutch datasets highlight the resulting ``empirical interplay" between accuracy, fairness, and privacy.

cs.LG↗

arXiv2021

Sleeping Combinatorial Bandits

In this paper, we study an interesting combination of sleeping and combinatorial stochastic bandits. In the mixed model studied here, at each discrete time instant, an arbitrary \emph{availability set} is generated from a fixed set of \emph{base} arms. An algorithm can select a subset of arms from the \emph{availability set} (sleeping bandits) and receive the corresponding reward along with semi-bandit feedback (combinatorial bandits). We adapt the well-known CUCB algorithm in the sleeping combinatorial bandits setting and refer to it as \CSUCB. We prove -- under mild smoothness conditions -- that the \CSUCB\ algorithm achieves an $O(\log (T))$ instance-dependent regret guarantee. We further prove that (i) when the range of the rewards is bounded, the regret guarantee of \CSUCB\ algorithm is $O(\sqrt{T \log (T)})$ and (ii) the instance-independent regret is $O(\sqrt[3]{T^2 \log(T)})$ in a general setting. Our results are quite general and hold under general environments -- such as non-additive reward functions, volatile arm availability, a variable number of base-arms to be pulled -- arising in practical applications. We validate the proven theoretical guarantees through experiments.

cs.LG↗

arXiv2021

Ballooning Multi-Armed Bandits

In this paper, we introduce Ballooning Multi-Armed Bandits (BL-MAB), a novel extension of the classical stochastic MAB model. In the BL-MAB model, the set of available arms grows (or balloons) over time. In contrast to the classical MAB setting where the regret is computed with respect to the best arm overall, the regret in a BL-MAB setting is computed with respect to the best available arm at each time. We first observe that the existing stochastic MAB algorithms result in linear regret for the BL-MAB model. We prove that, if the best arm is equally likely to arrive at any time instant, a sub-linear regret cannot be achieved. Next, we show that if the best arm is more likely to arrive in the early rounds, one can achieve sub-linear regret. Our proposed algorithm determines (1) the fraction of the time horizon for which the newly arriving arms should be explored and (2) the sequence of arm pulls in the exploitation phase from among the explored arms. Making reasonable assumptions on the arrival distribution of the best arm in terms of the thinness of the distribution's tail, we prove that the proposed algorithm achieves sub-linear instance-independent regret. We further quantify explicit dependence of regret on the arrival distribution parameters. We reinforce our theoretical findings with extensive simulation results. We conclude by showing that our algorithm would achieve sub-linear regret even if (a) the distributional parameters are not exactly known, but are obtained using a reasonable learning mechanism or (b) the best arm is not more likely to arrive early, but a large fraction of arms is likely to arrive relatively early.

cs.LG↗

arXiv2021

FASTEN: Fair and Secure Distributed Voting Using Smart Contracts

Electing democratic representatives via voting has been a common mechanism since the 17th century. However, these mechanisms raise concerns about fairness, privacy, vote concealment, fair calculations of tally, and proxies voting on their behalf for the voters. Ballot voting, and in recent times, electronic voting via electronic voting machines (EVMs) improves fairness by relying on centralized trust. Homomorphic encryption-based voting protocols also assure fairness but cannot scale to large scale elections such as presidential elections. In this paper, we leverage the blockchain technology of distributing trust to propose a smart contract-based protocol, namely, \proto. There are many existing protocols for voting using smart contracts. We observe that these either are not scalable or leak the vote tally during the voting stage, i.e., do not provide vote concealment. In contrast, we show that FASTEN preserves voter's privacy ensures vote concealment, immutability, and avoids double voting. We prove that the probability of privacy breaches is negligibly small. Further, our cost analysis of executing FASTEN over Ethereum is comparable to most of the existing cost of elections.

cs.CR↗

arXiv2021

We might walk together, but I run faster: Network Fairness and Scalability in Blockchains

Blockchain-based Distributed Ledgers (DLs) promise to transform the existing financial system by making it truly democratic. In the past decade, blockchain technology has seen many novel applications ranging from the banking industry to real estate. However, in order to be adopted universally, blockchain systems must be scalable to support a high volume of transactions. As we increase the throughput of the DL system, the underlying peer-to-peer network might face multiple levels of challenges to keep up with the requirements. Due to varying network capacities, the slower nodes would be at a relative disadvantage compared to the faster ones, which could negatively impact their revenue. In order to quantify their relative advantage or disadvantage, we introduce two measures of network fairness, $p_f$, the probability of frontrunning and $α_f$, the publishing fairness. We show that as we scale the blockchain, both these measures deteriorate, implying that the slower nodes face a disadvantage at higher throughputs. It results in the faster nodes getting more than their fair share of the reward while the slower nodes (slow in terms of network quality) get less. Thus, fairness and scalability in blockchain systems do not go hand in hand. In a setting with rational miners, lack of fairness causes miners to deviate from the "longest chain rule" or undercut, which would reduce the blockchain's resilience against byzantine adversaries. Hence, fairness is not only a desirable property for a blockchain system but also essential for the security of the blockchain and any scalable blockchain protocol proposed must ensure fairness.

cs.DC↗

arXiv2021

A Multi-Arm Bandit Approach To Subset Selection Under Constraints

We explore the class of problems where a central planner needs to select a subset of agents, each with different quality and cost. The planner wants to maximize its utility while ensuring that the average quality of the selected agents is above a certain threshold. When the agents' quality is known, we formulate our problem as an integer linear program (ILP) and propose a deterministic algorithm, namely \dpss\ that provides an exact solution to our ILP. We then consider the setting when the qualities of the agents are unknown. We model this as a Multi-Arm Bandit (MAB) problem and propose \newalgo\ to learn the qualities over multiple rounds. We show that after a certain number of rounds, $τ$, \newalgo\ outputs a subset of agents that satisfy the average quality constraint with a high probability. Next, we provide bounds on $τ$ and prove that after $τ$ rounds, the algorithm incurs a regret of $O(\ln T)$, where $T$ is the total number of rounds. We further illustrate the efficacy of \newalgo\ through simulations. To overcome the computational limitations of \dpss, we propose a polynomial-time greedy algorithm, namely \greedy, that provides an approximate solution to our ILP. We also compare the performance of \dpss\ and \greedy\ through experiments.

cs.LG↗

arXiv2021

Towards Mobile Distributed Ledgers

Advances in mobile computing have paved the way for new types of distributed applications that can be executed solely by mobile devices on device-to-device (D2D) ecosystems (e.g., crowdsensing). Sophisticated applications, like cryptocurrencies, need distributed ledgers to function. Distributed ledgers, such as blockchains and directed acyclic graphs (DAGs), employ consensus protocols to add data in the form of blocks. However, such protocols are designed for resourceful devices that are interconnected via the Internet. Moreover, existing distributed ledgers are not deployable to D2D ecosystems since their storage needs are continuously increasing. In this work, we introduce and analyse Mneme, a DAG-based distributed ledger that can be maintained solely by mobile devices. Mneme utilizes two novel consensus protocols: Proof-of-Context (PoC) and Proof-of-Equivalence (PoE). PoC employs users' context to add data on Mneme. PoE is executed periodically to summarize data and produce equivalent blocks that require less storage. We analyze Mneme's security and justify the ability of PoC and PoE to guarantee the characteristics of distributed ledgers: persistence and liveness. Furthermore, we analyze potential attacks from malicious users and prove that the probability of a successful attack is inversely proportional to the square of the number of mobile users who maintain Mneme.

cs.DC↗

arXiv2020

FNNC: Achieving Fairness through Neural Networks

In classification models fairness can be ensured by solving a constrained optimization problem. We focus on fairness constraints like Disparate Impact, Demographic Parity, and Equalized Odds, which are non-decomposable and non-convex. Researchers define convex surrogates of the constraints and then apply convex optimization frameworks to obtain fair classifiers. Surrogates serve only as an upper bound to the actual constraints, and convexifying fairness constraints might be challenging. We propose a neural network-based framework, \emph{FNNC}, to achieve fairness while maintaining high accuracy in classification. The above fairness constraints are included in the loss using Lagrangian multipliers. We prove bounds on generalization errors for the constrained losses which asymptotically go to zero. The network is optimized using two-step mini-batch stochastic gradient descent. Our experiments show that FNNC performs as good as the state of the art, if not better. The experimental evidence supplements our theoretical guarantees. In summary, we have an automated solution to achieve fairness in classification, which is easily extendable to many fairness constraints.

cs.LG↗