arXiv subjects

Sukanya Kudva

Publications and source records attributed to Sukanya Kudva.

5 recordsLinked to original sources

arXiv2026

Collusion-proof Auction Design using Side Information

Existing auction mechanisms are vulnerable to bidder collusion, which substantially degrades revenue and non-colluder welfare. To design truthful mechanisms resilient to collusion, we introduce a novel approach that leverages a machine learning classifier to predict (even imprecisely) which bidders are colluding. We first establish a Bulow-Klemperer-type result for multi-unit auctions with single-minded bidders, demonstrating that collusion significantly harms existing mechanisms only when the colluding coalition is large. Consequently, we focus our design on settings with many colluders. Building on the welfare-optimal Vickrey-Clarke-Groves (VCG) mechanism, we propose two novel truthful mechanisms: VCG-Posted Price (V-PoP) and Conditional-Posted Price (C-PoP). V-PoP applies VCG to non-colluding bidders and posted prices to colluding ones, and ensuring truthfulness is non-trivial because we must dynamically split the quantity of items between these groups based on the values of the non-colluder bids. C-PoP further advances this by computing a posted price conditioned on non-colluder bids, and ensuring truthfulness is non-obvious because the posted price is chosen using the values of the non-colluder bids. Because real-world classifiers make errors, we provide theoretical lower bounds on the auction price of V-PoP and C-PoP under misclassification, which theory shows acts as a proxy for welfare and revenue. Crucially, our bounds yield actionable insights for classifier design, revealing that false negatives (misclassifying colluders as non-colluders) are preferable to false positives (misclassifying non-colluders as colluders). Numerical experiments demonstrate that our mechanisms achieve high welfare and revenue against collusion, even when utilizing simple, low-cost classifiers.

cs.GT↗

arXiv2024

Understanding the Impact of Coalitions between EV Charging Stations

The rapid growth of electric vehicles (EVs) is driving the expansion of charging infrastructure globally. As charging stations become ubiquitous, their substantial electricity consumption can influence grid operation and electricity pricing. Naturally, \textit{some} groups of charging stations, which could be jointly operated by a company, may coordinate to decide their charging profile. While coordination among all charging stations is ideal, it is unclear if coordination of some charging stations is better than no coordination. In this paper, we analyze this intermediate regime between no and full coordination of charging stations. We model EV charging as a non-cooperative aggregative game, where each station's cost is determined by both monetary payments tied to reactive electricity prices on the grid and its sensitivity to deviations from a desired charging profile. We consider a solution concept that we call $\mathcal{C}$-Nash equilibrium, which is tied to a coalition $\mathcal{C}$ of charging stations coordinating to reduce their costs. We provide sufficient conditions, in terms of the demand and sensitivity of charging stations, to determine when independent (aka uncoordinated) operation of charging stations could result in lower overall costs to charging stations, coalition and charging stations outside the coalition. Somewhat counter to common intuition, we show numerical instances where allowing charging stations to operate independently is better than coordinating a subset of stations as a coalition. Jointly, these results provide operators of charging stations insights into how to coordinate their charging behavior, and open several research directions.

math.OC↗

arXiv2024

Tensor Completion via Integer Optimization

The main challenge with the tensor completion problem is a fundamental tension between computation power and the information-theoretic sample complexity rate. Past approaches either achieve the information-theoretic rate but lack practical algorithms to compute the corresponding solution, or have polynomial-time algorithms that require an exponentially-larger number of samples for low estimation error. This paper develops a novel tensor completion algorithm that resolves this tension by achieving both provable convergence (in numerical tolerance) in a linear number of oracle steps and the information-theoretic rate. Our approach formulates tensor completion as a convex optimization problem constrained using a gauge-based tensor norm, which is defined in a way that allows the use of integer linear optimization to solve linear separation problems over the unit-ball in this new norm. Adaptations based on this insight are incorporated into a Frank-Wolfe variant to build our algorithm. We show our algorithm scales-well using numerical experiments on tensors with up to ten million entries.

math.OC↗

arXiv2022

When would online platforms pay data dividends

Online platforms, including social media and search platforms, have routinely used their users' data for targeted ads, to improve their services, and to sell to third-party buyers. But an increasing awareness of the importance of users' data privacy has led to new laws that regulate data-sharing by platforms. Further, there have been political discussions on introducing data dividends, that is paying users for their data. Three interesting questions are then: When would these online platforms be incentivized to pay data dividends? How does their decision depend on whether users value their privacy more than the platform's free services? And should platforms invest in protecting users' data? This paper considers various factors affecting the users' and platform's decisions through utility functions. We construct a principal-agent model using a Stackelberg game to calculate their optimal decisions and qualitatively discuss the implications. Our results could inform a policymaker trying to understand the consequences of mandating data dividends.

cs.GT↗

arXiv2019

Few-electrode design for silicon MOS quantum dots

Silicon metal-oxide-semiconductor (MOS) spin qubits have become a promising platform for quantum information processing, with recent demonstrations of high-fidelity single and two-qubit gates. To move beyond a few qubits, however, more scalable designs that reduce the fabrication complexity and electrode density are needed. Here, we introduce a two-metal-layer MOS quantum dot device in which tunnel barriers are naturally formed by gaps between electrodes and controlled by adjacent accumulation gates. The accumulation gates define the electron reservoirs and provide tunability of the tunnel rate of nearly 8.5 decades/V, determined by a combination of charge sensor electron counting measurements and by direct transport. The valley splitting in the few-electron regime is probed by magneto-spectroscopy up to a field of 6 T, providing an estimate for the ground-state gap of 290 $μ$eV. We show preliminary characterization of a double quantum dot, demonstrating that this design can be extended to linear dot arrays that should be useful in applications like electron shuttling. These results motivate further innovations in MOS quantum dot design that can improve the scalability prospects for spin qubits.

cond-mat.mes-hall↗