arXiv subjects

Shuchi Chawla

Publications and source records attributed to Shuchi Chawla.

At least 37 records · Page 2Linked to original sources

arXiv2022

Individually-Fair Auctions for Multi-Slot Sponsored Search

We design fair sponsored search auctions that achieve a near-optimal tradeoff between fairness and quality. Our work builds upon the model and auction design of Chawla and Jagadeesan \cite{CJ22}, who considered the special case of a single slot. We consider sponsored search settings with multiple slots and the standard model of click through rates that are multiplicatively separable into an advertiser-specific component and a slot-specific component. When similar users have similar advertiser-specific click through rates, our auctions achieve the same near-optimal tradeoff between fairness and quality as in \cite{CJ22}. When similar users can have different advertiser-specific preferences, we show that a preference-based fairness guarantee holds. Finally, we provide a computationally efficient algorithm for computing payments for our auctions as well as those in previous work, resolving another open direction from \cite{CJ22}.

cs.GT↗

arXiv2022

Cite-seeing and Reviewing: A Study on Citation Bias in Peer Review

Citations play an important role in researchers' careers as a key factor in evaluation of scientific impact. Many anecdotes advice authors to exploit this fact and cite prospective reviewers to try obtaining a more positive evaluation for their submission. In this work, we investigate if such a citation bias actually exists: Does the citation of a reviewer's own work in a submission cause them to be positively biased towards the submission? In conjunction with the review process of two flagship conferences in machine learning and algorithmic economics, we execute an observational study to test for citation bias in peer review. In our analysis, we carefully account for various confounding factors such as paper quality and reviewer expertise, and apply different modeling techniques to alleviate concerns regarding the model mismatch. Overall, our analysis involves 1,314 papers and 1,717 reviewers and detects citation bias in both venues we consider. In terms of the effect size, by citing a reviewer's work, a submission has a non-trivial chance of getting a higher score from the reviewer: an expected increase in the score is approximately 0.23 on a 5-point Likert item. For reference, a one-point increase of a score by a single reviewer improves the position of a submission by 11% on average.

cs.DL↗

arXiv2022

Mechanism Redesign

This paper develops the theory of mechanism redesign by which an auctioneer can reoptimize an auction based on bid data collected from previous iterations of the auction on bidders from the same market. We give a direct method for estimation of the revenue of a counterfactual auction from the bids in the current auction. The estimator is a simple weighted order statistic of the bids and has the optimal error rate. Two applications of our estimator are A/B testing (a.k.a., randomized controlled trials) and instrumented optimization (i.e., revenue optimization subject to being able to do accurate inference of any counterfactual auction revenue).

cs.GT↗

arXiv2021

Individual Fairness in Advertising Auctions through Inverse Proportionality

Recent empirical work demonstrates that online advertisement can exhibit bias in the delivery of ads across users even when all advertisers bid in a non-discriminatory manner. We study the design of ad auctions that, given fair bids, are guaranteed to produce fair outcomes. Following the works of Dwork and Ilvento (2019) and Chawla et al. (2020), our goal is to design a truthful auction that satisfies ``individual fairness'' in its outcomes: informally speaking, users that are similar to each other should obtain similar allocations of ads. Within this framework we quantify the tradeoff between social welfare maximization and fairness. This work makes two conceptual contributions. First, we express the fairness constraint as a kind of stability condition: any two users that are assigned multiplicatively similar values by all the advertisers must receive additively similar allocations for each advertiser. This value stability constraint is expressed as a function that maps the multiplicative distance between value vectors to the maximum allowable $\ell_{\infty}$ distance between the corresponding allocations. Standard auctions do not satisfy this kind of value stability. Second, we introduce a new class of allocation algorithms called Inverse Proportional Allocation that achieve a near optimal tradeoff between fairness and social welfare for a broad and expressive class of value stability conditions. These allocation algorithms are truthful and prior-free, and achieve a constant factor approximation to the optimal (unconstrained) social welfare. In particular, the approximation ratio is independent of the number of advertisers in the system. In this respect, these allocation algorithms greatly surpass the guarantees achieved in previous work. We also extend our results to broader notions of fairness that we call subset fairness.

cs.GT↗

arXiv2021

Pricing Ordered Items

We study the revenue guarantees and approximability of item pricing. Recent work shows that with $n$ heterogeneous items, item-pricing guarantees an $O(\log n)$ approximation to the optimal revenue achievable by any (buy-many) mechanism, even when buyers have arbitrarily combinatorial valuations. However, finding good item prices is challenging -- it is known that even under unit-demand valuations, it is NP-hard to find item prices that approximate the revenue of the optimal item pricing better than $O(\sqrt{n})$. Our work provides a more fine-grained analysis of the revenue guarantees and computational complexity in terms of the number of item ``categories'' which may be significantly fewer than $n$. We assume the items are partitioned in $k$ categories so that items within a category are totally-ordered and a buyer's value for a bundle depends only on the best item contained from every category. We show that item-pricing guarantees an $O(\log k)$ approximation to the optimal (buy-many) revenue and provide a PTAS for computing the optimal item-pricing when $k$ is constant. We also provide a matching lower bound showing that the problem is (strongly) NP-hard even when $k=1$. Our results naturally extend to the case where items are only partially ordered, in which case the revenue guarantees and computational complexity depend on the width of the partial ordering, i.e. the largest set for which no two items are comparable.

cs.GT↗

arXiv2021

Visions in Theoretical Computer Science: A Report on the TCS Visioning Workshop 2020

Theoretical computer science (TCS) is a subdiscipline of computer science that studies the mathematical foundations of computational and algorithmic processes and interactions. Work in this field is often recognized by its emphasis on mathematical technique and rigor. At the heart of the field are questions surrounding the nature of computation: What does it mean to compute? What is computable? And how efficiently? Every ten years or so the TCS community attends visioning workshops to discuss the challenges and recent accomplishments in the TCS field. The workshops and the outputs they produce are meant both as a reflection for the TCS community and as guiding principles for interested investment partners. Concretely, the workshop output consists of a number of nuggets, each summarizing a particular point, that are synthesized in the form of a white paper and illustrated with graphics/slides produced by a professional graphic designer. The second TCS Visioning Workshop was organized by the SIGACT Committee for the Advancement of Theoretical Computer Science and took place during the week of July 20, 2020. Despite the conference being virtual, there were over 76 participants, mostly from the United States, but also a few from Europe and Asia who were able to attend due to the online format. Workshop participants were divided into categories as reflected in the sections of this report: (1) models of computation; (2) foundations of data science; (3) cryptography; and (4) using theoretical computer science for other domains. Each group participated in a series of discussions that produced the nuggets below.

cs.CY↗

arXiv2020

Non-Adaptive Matroid Prophet Inequalities

We investigate non-adaptive algorithms for matroid prophet inequalities. Matroid prophet inequalities have been considered resolved since 2012 when [KW12] introduced thresholds that guarantee a tight 2-approximation to the prophet; however, this algorithm is adaptive. Other approaches of [CHMS10] and [FSZ16] have used non-adaptive thresholds with a feasibility restriction; however, this translates to adaptively changing an item's threshold to infinity when it cannot be taken with respect to the additional feasibility constraint, hence the algorithm is not truly non-adaptive. A major application of prophet inequalities is in auction design, where non-adaptive prices possess a significant advantage: they convert to order-oblivious posted pricings, and are essential for translating a prophet inequality into a truthful mechanism for multi-dimensional buyers. The existing matroid prophet inequalities do not suffice for this application. We present the first non-adaptive constant-factor prophet inequality for graphic matroids.

cs.DS↗

arXiv2020

Pandora's Box with Correlations: Learning and Approximation

The Pandora's Box problem and its extensions capture optimization problems with stochastic input where the algorithm can obtain instantiations of input random variables at some cost. To our knowledge, all previous work on this class of problems assumes that different random variables in the input are distributed independently. As such it does not capture many real-world settings. In this paper, we provide the first approximation algorithms for Pandora's Box-type problems with correlations. We assume that the algorithm has access to samples drawn from the joint distribution on input. Algorithms for these problems must determine an order in which to probe random variables, as well as when to stop and return the best solution found so far. In general, an optimal algorithm may make both decisions adaptively based on instantiations observed previously. Such fully adaptive (FA) strategies cannot be efficiently approximated to within any sublinear factor with sample access. We therefore focus on the simpler objective of approximating partially adaptive (PA) strategies that probe random variables in a fixed predetermined order but decide when to stop based on the instantiations observed. We consider a number of different feasibility constraints and provide simple PA strategies that are approximately optimal with respect to the best PA strategy for each case. All of our algorithms have polynomial sample complexity. We further show that our results are tight within constant factors: better factors cannot be achieved even using the full power of FA strategies.

cs.DS↗

arXiv2020

Menu-size Complexity and Revenue Continuity of Buy-many Mechanisms

We study the multi-item mechanism design problem where a monopolist sells $n$ heterogeneous items to a single buyer. We focus on buy-many mechanisms, a natural class of mechanisms frequently used in practice. The buy-many property allows the buyer to interact with the mechanism multiple times instead of once as in the more commonly studied buy-one mechanisms. This imposes additional incentive constraints and thus restricts the space of mechanisms that the seller can use. In this paper, we explore the qualitative differences between buy-one and buy-many mechanisms focusing on two important properties: revenue continuity and menu-size complexity. Our first main result is that when the value function of the buyer is perturbed multiplicatively by a factor of $1\pmε$, the optimal revenue obtained by buy-many mechanisms only changes by a factor of $1 \pm \textrm{poly}(n,ε)$. In contrast, for buy-one mechanisms, the revenue of the resulting optimal mechanism after such a perturbation can change arbitrarily. Our second main result shows that under any distribution of arbitrary valuations, finite menu size suffices to achieve a $(1-ε)$-approximation to the optimal buy-many mechanism. We give tight upper and lower bounds on the number of menu entries as a function of $n$ and $ε$. On the other hand, such a result fails to hold for buy-one mechanisms as even for two items and a buyer with either unit-demand or additive valuations, the menu-size complexity of approximately optimal mechanisms is unbounded.

cs.GT↗

arXiv2019

Themis: Fair and Efficient GPU Cluster Scheduling

Modern distributed machine learning (ML) training workloads benefit significantly from leveraging GPUs. However, significant contention ensues when multiple such workloads are run atop a shared cluster of GPUs. A key question is how to fairly apportion GPUs across workloads. We find that established cluster scheduling disciplines are a poor fit because of ML workloads' unique attributes: ML jobs have long-running tasks that need to be gang-scheduled, and their performance is sensitive to tasks' relative placement. We propose Themis, a new scheduling framework for ML training workloads. It's GPU allocation policy enforces that ML workloads complete in a finish-time fair manner, a new notion we introduce. To capture placement sensitivity and ensure efficiency, Themis uses a two-level scheduling architecture where ML workloads bid on available resources that are offered in an auction run by a central arbiter. Our auction design allocates GPUs to winning bids by trading off efficiency for fairness in the short term but ensuring finish-time fairness in the long term. Our evaluation on a production trace shows that Themis can improve fairness by more than 2.25X and is ~5% to 250% more cluster efficient in comparison to state-of-the-art schedulers.

cs.DC↗

arXiv2019

Revenue Maximization for Query Pricing

Buying and selling of data online has increased substantially over the last few years. Several frameworks have already been proposed that study query pricing in theory and practice. The key guiding principle in these works is the notion of {\em arbitrage-freeness} where the broker can set different prices for different queries made to the dataset, but must ensure that the pricing function does not provide the buyers with opportunities for arbitrage. However, little is known about revenue maximization aspect of query pricing. In this paper, we study the problem faced by a broker selling access to data with the goal of maximizing her revenue. We show that this problem can be formulated as a revenue maximization problem with single-minded buyers and unlimited supply, for which several approximation algorithms are known. We perform an extensive empirical evaluation of the performance of several pricing algorithms for the query pricing problem on real-world instances. In addition to previously known approximation algorithms, we propose several new heuristics and analyze them both theoretically and experimentally. Our experiments show that algorithms with the best theoretical bounds are not necessarily the best empirically. We identify algorithms and heuristics that are both fast and also provide consistently good performance when valuations are drawn from a variety of distributions.

cs.DB↗

arXiv2019

Multi-Category Fairness in Sponsored Search Auctions

Fairness in advertising is a topic of particular concern motivated by theoretical and empirical observations in both the computer science and economics literature. We examine the problem of fairness in advertising for general purpose platforms that service advertisers from many different categories. First, we propose inter-category and intra-category fairness desiderata that take inspiration from individual fairness and envy-freeness. Second, we investigate the "platform utility" (a proxy for the quality of the allocation) achievable by mechanisms satisfying these desiderata. More specifically, we compare the utility of fair mechanisms against the unfair optimal, and we show by construction that our fairness desiderata are compatible with utility. That is, we construct a family of fair mechanisms with high utility that perform close to optimally within a class of fair mechanisms. Our mechanisms also enjoy nice implementation properties including metric-obliviousness, which allows the platform to produce fair allocations without needing to know the specifics of the fairness requirements.

cs.GT↗

arXiv2019

Buy-many mechanisms are not much better than item pricing

Multi-item mechanisms can be very complex offering many different bundles to the buyer that could even be randomized. Such complexity is thought to be necessary as the revenue gaps between randomized and deterministic mechanisms, or deterministic and simple mechanisms are huge even for additive valuations. We challenge this conventional belief by showing that these large gaps can only happen in restricted situations. These are situations where the mechanism overcharges a buyer for a bundle while selling individual items at much lower prices. Arguably this is impractical in many settings because the buyer can break his order into smaller pieces paying a much lower price overall. Our main result is that if the buyer is allowed to purchase as many (randomized) bundles as he pleases, the revenue of any multi-item mechanism is at most O(logn) times the revenue achievable by item pricing, where n is the number of items. This holds in the most general setting possible, with an arbitrarily correlated distribution of buyer types and arbitrary valuations. We also show that this result is tight in a very strong sense. Any family of mechanisms of subexponential description complexity cannot achieve better than logarithmic approximation even against the best deterministic mechanism and even for additive valuations. In contrast, item pricing that has linear description complexity matches this bound against randomized mechanisms.

cs.GT↗

arXiv2018

Pricing for Online Resource Allocation: Intervals and Paths

We present pricing mechanisms for several online resource allocation problems which obtain tight or nearly tight approximations to social welfare. In our settings, buyers arrive online and purchase bundles of items; buyers' values for the bundles are drawn from known distributions. This problem is closely related to the so-called prophet-inequality of Krengel and Sucheston and its extensions in recent literature. Motivated by applications to cloud economics, we consider two kinds of buyer preferences. In the first, items correspond to different units of time at which a resource is available; the items are arranged in a total order and buyers desire intervals of items. The second corresponds to bandwidth allocation over a tree network; the items are edges in the network and buyers desire paths. Because buyers' preferences have complementarities in the settings we consider, recent constant-factor approximations via item prices do not apply, and indeed strong negative results are known. We develop static, anonymous bundle pricing mechanisms. For the interval preferences setting, we show that static, anonymous bundle pricings achieve a sublogarithmic competitive ratio, which is optimal (within constant factors) over the class of all online allocation algorithms, truthful or not. For the path preferences setting, we obtain a nearly-tight logarithmic competitive ratio. Both of these results exhibit an exponential improvement over item pricings for these settings. Our results extend to settings where the seller has multiple copies of each item, with the competitive ratio decreasing linearly with supply. Such a gradual tradeoff between supply and the competitive ratio for welfare was previously known only for the single item prophet inequality.

cs.GT↗

arXiv2018

Revenue Maximization with an Uncertainty-Averse Buyer

Most work in mechanism design assumes that buyers are risk neutral; some considers risk aversion arising due to a non-linear utility for money. Yet behavioral studies have established that real agents exhibit risk attitudes which cannot be captured by any expected utility model. We initiate the study of revenue-optimal mechanisms under buyer behavioral models beyond expected utility theory. We adopt a model from prospect theory which arose to explain these discrepancies and incorporates agents under-weighting uncertain outcomes. In our model, an event occurring with probability $x < 1$ is worth strictly less to the agent than $x$ times the value of the event when it occurs with certainty. In contrast to the risk-neutral setting, the optimal mechanism may be randomized and appears challenging to find, even for a single buyer and a single item for sale. Nevertheless, we give a characterization of the optimal mechanism which enables positive approximation results. In particular, we show that under a reasonable bounded-risk-aversion assumption, posted pricing obtains a constant approximation. Notably, this result is "risk-robust" in that it does not depend on the details of the buyer's risk attitude. Finally, we examine a dynamic setting in which the buyer is uncertain about his future value. In contrast to positive results for a risk-neutral buyer, we show that the buyer's risk aversion may prevent the seller from approximating the optimal revenue in a risk-robust manner.

cs.GT↗

arXiv2017

Stability of Service under Time-of-Use Pricing

We consider "time-of-use" pricing as a technique for matching supply and demand of temporal resources with the goal of maximizing social welfare. Relevant examples include energy, computing resources on a cloud computing platform, and charging stations for electric vehicles, among many others. A client/job in this setting has a window of time during which he needs service, and a particular value for obtaining it. We assume a stochastic model for demand, where each job materializes with some probability via an independent Bernoulli trial. Given a per-time-unit pricing of resources, any realized job will first try to get served by the cheapest available resource in its window and, failing that, will try to find service at the next cheapest available resource, and so on. Thus, the natural stochastic fluctuations in demand have the potential to lead to cascading overload events. Our main result shows that setting prices so as to optimally handle the {\em expected} demand works well: with high probability, when the actual demand is instantiated, the system is stable and the expected value of the jobs served is very close to that of the optimal offline algorithm.

cs.GT↗

arXiv2017

Truth and Regret in Online Scheduling

We consider a scheduling problem where a cloud service provider has multiple units of a resource available over time. Selfish clients submit jobs, each with an arrival time, deadline, length, and value. The service provider's goal is to implement a truthful online mechanism for scheduling jobs so as to maximize the social welfare of the schedule. Recent work shows that under a stochastic assumption on job arrivals, there is a single-parameter family of mechanisms that achieves near-optimal social welfare. We show that given any such family of near-optimal online mechanisms, there exists an online mechanism that in the worst case performs nearly as well as the best of the given mechanisms. Our mechanism is truthful whenever the mechanisms in the given family are truthful and prompt, and achieves optimal (within constant factors) regret. We model the problem of competing against a family of online scheduling mechanisms as one of learning from expert advice. A primary challenge is that any scheduling decisions we make affect not only the payoff at the current step, but also the resource availability and payoffs in future steps. Furthermore, switching from one algorithm (a.k.a. expert) to another in an online fashion is challenging both because it requires synchronization with the state of the latter algorithm as well as because it affects the incentive structure of the algorithms. We further show how to adapt our algorithm to a non-clairvoyant setting where job lengths are unknown until jobs are run to completion. Once again, in this setting, we obtain truthfulness along with asymptotically optimal regret (within poly-logarithmic factors).

cs.GT↗

arXiv2016

Timing Matters: Online Dynamics in Broadcast Games

A central question in algorithmic game theory is to measure the inefficiency (ratio of costs) of Nash equilibria (NE) with respect to socially optimal solutions. The two established metrics used for this purpose are price of anarchy (POA) and price of stability (POS), which respectively provide upper and lower bounds on this ratio. A deficiency of these metrics, however, is that they are purely existential and shed no light on which of the equilibrium states are reachable in an actual game, i.e., via natural game dynamics. This is particularly striking if these metrics differ significantly, such as in network design games where the exponential gap between the best and worst NE states originally prompted the notion of POS in game theory (Anshelevich et al., FOCS 2002). In this paper, we make progress toward bridging this gap by studying network design games under natural game dynamics. First we show that in a completely decentralized setting, where agents arrive, depart, and make improving moves in an arbitrary order, the inefficiency of NE attained can be polynomially large. This implies that the game designer must have some control over the interleaving of these events in order to force the game to attain efficient NE. We complement our negative result by showing that if the game designer is allowed to execute a sequence of improving moves to create an equilibrium state after every batch of agent arrivals or departures, then the resulting equilibrium states attained by the game are exponentially more efficient, i.e., the ratio of costs compared to the optimum is only logarithmic. Overall, our two results establish that in network games, the efficiency of equilibrium states is dictated by whether agents are allowed to join or leave the game in arbitrary states, an observation that might be useful in analyzing the dynamics of other classes of games with divergent POS and POA bounds.

cs.DS↗