SearcharxivSearch

arXiv subjects

Eva Tardos

Publications and source records attributed to Eva Tardos.

At least 19 recordsLinked to original sources

Improved Multi-Dimensional Forecasting for Swap Regret

We study the problem of forecasting for an arbitrary number of downstream agents with unknown objectives, each of whom best responds to the forecaster's predictions. We seek a single forecaster that guarantees sublinear swap regret for all downstream agents simultaneously. For two-dimensional outcome spaces, we give a polynomial time algorithm that guarantees $\tilde{O}(\sqrt{kT})$ swap regret for any downstream agent with $k$ actions. This improves over the previously known bound of $\tilde{O}(kT^{5/8})$ and avoids the exponential in $T$ runtime of prior algorithms in this setting. Our algorithm extends nicely to other low dimensional environments, retaining $\tilde{O}(\sqrt{T})$ downstream swap regret while the exponent of $k$ in the regret bound and the exponent of $T$ in the running time both grow with dimension. For arbitrary dimension $d$, we give a forecasting algorithm that guarantees $\tilde{O}(d\sqrt{kT})$ swap regret, assuming the forecaster knows an upper bound $k$ on the number of actions available to any downstream agent, albeit with a much longer runtime. This improves upon previous high dimensional guarantees that had $\tilde{O}(T^{2/3})$ dependence and required additional behavioral assumptions.

cs.GT

Quantifying Theoretical AI Alignment Guarantees: Receiver-Utility Bounds in Bayesian Persuasion

Misalignment can change how information moves from an AI agent to a human user. We model this as an information advantage: the AI agent observes the world state, while the human receiver only knows a prior and must act after seeing the agent's signal. A strategic AI sender may withhold evidence or garble information in order to steer the human's decision. We ask how much useful information can still reach the human when the AI optimizes a misaligned objective. We study a Bayesian persuasion model in which the world state is a bit string, the human receiver wants to guess the bits correctly, and a single AI sender wants the receiver to guess as many bits as possible as $1$. For a prior $\mu$, let $R_0(\mu)$ be the receiver's utility from using only the prior, and let $R_{\max}(\mu)$ be the largest receiver utility among signaling schemes that are optimal for the sender. We prove $R_{\max}(\mu)/R_0(\mu)\leq 3/2$. This bound improves for priors close to the independent product prior with the same marginals: if $\mu(x)\geq (1-\eta)\pi_\mu(x)$ for every state $x$, then $R_{\max}(\mu)\leq R_0(\mu)+\eta n$. We also give a six-bit prior for which $R_{\max}(\mu)/R_0(\mu)=39/31>5/4$, so no universal $5/4$ bound is possible.

cs.GT

Robust Temporal Guarantees in Budgeted Sequential Auctions

In modern advertising platforms, learning algorithms are deployed by budget-constrained bidders to maximize their accumulated value. These algorithms often offer classical utility guarantees like no-regret, i.e., the agent's utility is at least the utility achieved by some benchmark in which it is assumed that every other agent's bidding remains the same. These guarantees offer compelling properties: They are optimal against stationary competition distributions, and in unconstrained settings, the resulting empirical distribution of play induced by no-regret dynamics approximates a Coarse Correlated Equilibrium. However, no-regret algorithms are easily manipulable, and in budgeted settings, no stronger notion of regret (such as swap regret) is currently known that would limit such manipulation. We propose a very simple learning algorithm for budgeted sequential auctions where agents maximize their total number of wins and show that it has surprisingly appealing properties. We analyze this algorithm from two perspectives. First, we show that when an agent with a $\rho$ fraction of the total budget uses this algorithm, then she is guaranteed to win at least $\rho T - O(\sqrt T)$ of the total $T$ rounds. This result holds for adversarial behavior by the other agents, as long as they respect their own budget restrictions. Second, we examine the scenario when all the agents follow our algorithm. By the first result, every agent's total wins are proportional to her budget, up to the additive $O(\sqrt T)$ term. In addition, we show that this result holds in a much stronger sense: after an initial period of $O(\sqrt T \log T)$ rounds, every agent gets the same guarantee over any time interval. For intervals of length $O(\sqrt T)$, we show that the deviation from the desired number of wins is an additive constant.

cs.GT

Stable and Fault-Tolerant Decentralized Traffic Engineering

Cloud providers have recently decentralized their wide-area network traffic engineering (TE) systems to contain the impact of TE controller failures. In the decentralized design, a controller fault only impacts its slice of the network, limiting the blast radius to a fraction of the network. However, we find that autonomous slice controllers can arrive at divergent traffic allocations that overload links by 30% beyond their capacity. We present Symphony, a decentralized TE system that addresses the challenge of divergence-induced congestion while preserving the fault-isolation benefits of decentralization. By augmenting TE objectives with quadratic regularization, Symphony makes traffic allocations robust to demand perturbations, ensuring TE controllers naturally converge to compatible allocations without coordination. In parallel, Symphony's randomized slicing algorithm partitions the network to minimize blast radius by distributing critical traffic sources across slices, preventing any single failure from becoming catastrophic. These innovations work in tandem: regularization ensures algorithmic stability to traffic allocations while intelligent slicing provides architectural resilience in the network. Through extensive evaluation on cloud provider WANs, we show Symphony reduces divergence-induced congestion by 14x and blast radius by 79% compared to current practice.

cs.NI

Asymmetric Network Games: $\alpha$-Potential Function and Learning

In a network game, players interact over a network and the utility of each player depends on his own action and on an aggregate of his neighbours' actions. Many real world networks of interest are asymmetric and involve a large number of heterogeneous players. This paper analyzes static network games using the framework of $\alpha$-potential games. Under mild assumptions on the action sets (compact intervals) and the utility functions (twice continuously differentiable) of the players, we derive an expression for an inexact potential function of the game, called the $\alpha$-potential function. Using such a function, we show that modified versions of the sequential best-response algorithm and the simultaneous gradient play algorithm achieve convergence of players' actions to a $2\alpha$-Nash equilibrium. For linear-quadratic network games, we show that $\alpha$ depends on the maximum asymmetry in the network and is well-behaved for a wide range of networks of practical interest. Further, we derive bounds on the social welfare of the $\alpha$-Nash equilibrium corresponding to the maximum of the $\alpha$-potential function, under suitable assumptions. We numerically illustrate the convergence of the proposed algorithms and properties of the learned $2\alpha$-Nash equilibria.

cs.GT

Learning in Strategic Queuing Systems with Small Buffers

We consider learning outcomes in games with carryover effects between rounds: when outcomes in the present round affect the game in the future. An important example of such systems is routers in networking, as they use simple learning algorithms to find the best way to deliver packets to their desired destination. This simple, myopic, and distributed decision process makes large queuing systems easy to operate, but at the same time, the system needs more capacity than would be required if all traffic were centrally coordinated. Gaitonde and Tardos (EC 2020 and JACM 2023) initiated the study of such systems, modeling them as an infinitely repeated game in which routers compete for servers and the system maintains a state (the number of packets held at each queue) that results from outcomes of previous rounds. However, their model assumes that servers have no buffers at all, so routers have to resend all packets that were not served successfully, which makes their system model unrealistic. They show that in their model, even with hugely increased server capacity relative to what is needed in the centrally coordinated case, ensuring that the system is stable requires the use of timestamps and priority for older packets. We consider a system with two important changes, which make the model more realistic and allow for much higher traffic rates: first, we add a very small buffer to each server, allowing the server to hold on to a single packet to be served later (if it fails to serve it immediately), and second, we do not require timestamps or priority to older packets. Using theoretical analysis and simulations, we show that when queues are learning, a small constant-factor increase in server capacity, compared to what would be needed if centrally coordinating, suffices to keep the system stable, even if servers select randomly among packets arriving simultaneously.

cs.GT

Markets with Heterogeneous Agents: Dynamics and Survival of Bayesian vs. No-Regret Learners

We analyze the performance of heterogeneous learning agents in asset markets with stochastic payoffs. Our main focus is on comparing Bayesian learners and no-regret learners who compete in markets and identifying the conditions under which each approach is more effective. We formally relate the notions of survival and market dominance studied in economics and the framework of regret minimization, thereby bridging these theories. A central finding is that regret plays a key role in market selection, but low regret alone does not guarantee survival: surprisingly, an agent may achieve even logarithmic regret and yet be driven out of the market when competing against a Bayesian learner with a finite prior that assigns positive probability to the correct model. At the same time, we show that Bayesian learning is highly fragile, while no-regret learning requires less knowledge of the environment and is therefore more robust. Motivated by this contrast, we propose two simple hybrid strategies that incorporate Bayesian updates while improving robustness and adaptability to distribution shifts, taking a step toward a best-of-both-worlds learning approach. More broadly, our work contributes to the understanding of dynamics of heterogeneous learning agents and their impact on markets.

cs.GT

Karma: Resource Allocation for Dynamic Demands

We consider the problem of fair resource allocation in a system where user demands are dynamic, that is, where user demands vary over time. Our key observation is that the classical max-min fairness algorithm for resource allocation provides many desirable properties (e.g., Pareto efficiency, strategy-proofness, and fairness), but only under the strong assumption of user demands being static over time. For the realistic case of dynamic user demands, the max-min fairness algorithm loses one or more of these properties. We present Karma, a new resource allocation mechanism for dynamic user demands. The key technical contribution in Karma is a credit-based resource allocation algorithm: in each quantum, users donate their unused resources and are assigned credits when other users borrow these resources; Karma carefully orchestrates the exchange of credits across users (based on their instantaneous demands, donated resources and borrowed resources), and performs prioritized resource allocation based on users' credits. We theoretically establish Karma guarantees related to Pareto efficiency, strategy-proofness, and fairness for dynamic user demands. Empirical evaluations over production workloads show that these properties translate well into practice: Karma is able to reduce disparity in performance across users to a bare minimum while maintaining Pareto-optimal system-wide performance.

cs.OS

Dynamic Pricing Provides Robust Equilibria in Stochastic Ridesharing Networks

Ridesharing markets are complex: drivers are strategic, rider demand and driver availability are stochastic, and complex city-scale phenomena like weather induce large scale correlation across space and time. At the same time, past work has focused on a subset of these challenges. We propose a model of ridesharing networks with strategic drivers, spatiotemporal dynamics, and stochasticity. Supporting both computational tractability and better modeling flexibility than classical fluid limits, we use a two-level stochastic model that allows correlated shocks caused by weather or large public events. Using this model, we propose a novel pricing mechanism: stochastic spatiotemporal pricing (SSP). We show that the SSP mechanism is asymptotically incentive-compatible and that all (approximate) equilibria of the resulting game are asymptotically welfare-maximizing when the market is large enough. The SSP mechanism iteratively recomputes prices based on realized demand and supply, and in this sense prices dynamically. We show that this is critical: while a static variant of the SSP mechanism (whose prices vary with the market-level stochastic scenario but not individual rider and driver decisions) has a sequence of asymptotically welfare-optimal approximate equilibria, we demonstrate that it also has other equilibria producing extremely low social welfare. Thus, we argue that dynamic pricing is important for ensuring robustness in stochastic ride-sharing networks.

math.OC

Small-loss bounds for online learning with partial information

We consider the problem of adversarial (non-stochastic) online learning with partial information feedback, where at each round, a decision maker selects an action from a finite set of alternatives. We develop a black-box approach for such problems where the learner observes as feedback only losses of a subset of the actions that includes the selected action. When losses of actions are non-negative, under the graph-based feedback model introduced by Mannor and Shamir, we offer algorithms that attain the so called "small-loss" $o(αL^{\star})$ regret bounds with high probability, where $α$ is the independence number of the graph, and $L^{\star}$ is the loss of the best action. Prior to our work, there was no data-dependent guarantee for general feedback graphs even for pseudo-regret (without dependence on the number of actions, i.e. utilizing the increased information feedback). Taking advantage of the black-box nature of our technique, we extend our results to many other applications such as semi-bandits (including routing in networks), contextual bandits (even with an infinite comparator class), as well as learning with slowly changing (shifting) comparators. In the special case of classical bandit and semi-bandit problems, we provide optimal small-loss, high-probability guarantees of $\tilde{O}(\sqrt{dL^{\star}})$ for actual regret, where $d$ is the number of actions, answering open questions of Neu. Previous bounds for bandits and semi-bandits were known only for pseudo-regret and only in expectation. We also offer an optimal $\tilde{O}(\sqrt{κL^{\star}})$ regret guarantee for fixed feedback graphs with clique-partition number at most $κ$.

cs.LG

Virtues of Patience in Strategic Queuing Systems

We consider the problem of selfish agents in discrete-time queuing systems, where competitive queues try to get their packets served. In this model, a queue gets to send a packet each step to one of the servers, which will attempt to serve the oldest arriving packet, and unprocessed packets are returned to each queue. We model this as a repeated game where queues compete for the capacity of the servers, but where the state of the game evolves as the length of each queue varies, resulting in a highly dependent random process. Earlier work by the authors [EC'20] shows that with no-regret learners, the system needs twice the capacity as would be required in the coordinated setting to ensure queue lengths remain stable despite the selfish behavior of the queues. In this paper, we demonstrate that this way of evaluating outcomes is myopic: if more patient queues choose strategies that selfishly maximize their long-run success rate, stability can be ensured with just $\frac{e}{e-1}\approx 1.58$ times extra capacity, better than what is possible assuming the no-regret property. As these systems induce highly dependent processes, our analysis draws heavily on techniques from probability theory. Though these systems are random under any fixed policies by the queues, we show that, surprisingly, these systems have deterministic and explicit asymptotic behavior. We show that the asymptotic growth rates of queues can be written as a ratio of a submodular and modular function, which provides significant game-theoretic properties. Our equilibrium analysis then relies on a novel deformation argument towards a more analyzable solution that differs significantly from previous price of anarchy results. While the intermediate points will not be equilibria, this analytic structure will ensure that this deformation is monotonic along this continuous path.

cs.GT

Adversarial Perturbations of Opinion Dynamics in Networks

We study the connections between network structure, opinion dynamics, and an adversary's power to artificially induce disagreements. We approach these questions by extending models of opinion formation in the social sciences to represent scenarios, familiar from recent events, in which external actors seek to destabilize communities through sophisticated information warfare tactics via fake news and bots. In many instances, the intrinsic goals of these efforts are not necessarily to shift the overall sentiment of the network, but rather to induce discord. These perturbations diffuse via opinion dynamics on the underlying network, through mechanisms that have been analyzed and abstracted through work in computer science and the social sciences. We investigate the properties of such attacks, considering optimal strategies both for the adversary seeking to create disagreement and for the entities tasked with defending the network from attack. We show that for different formulations of these types of objectives, different regimes of the spectral structure of the network will limit the adversary's capacity to sow discord; this enables us to qualitatively describe which networks are most vulnerable against these perturbations. We then consider the algorithmic task of a network defender to mitigate these sorts of adversarial attacks by insulating nodes heterogeneously; we show that, by considering the geometry of this problem, this optimization task can be efficiently solved via convex programming. Finally, we generalize these results to allow for two network structures, where the opinion dynamics process and the measurement of disagreement become uncoupled, and determine how the adversary's power changes; for instance, this may arise when opinion dynamics are controlled an online community via social media, while disagreement is measured along "real-world" connections.

cs.DS

Learning and Efficiency in Games with Dynamic Population

We study the quality of outcomes in repeated games when the population of players is dynamically changing and participants use learning algorithms to adapt to the changing environment. Game theory classically considers Nash equilibria of one-shot games, while in practice many games are played repeatedly, and in such games players often use algorithmic tools to learn to play in the given environment. Most previous work on learning in repeated games assumes that the population playing the game is static over time. We analyze the efficiency of repeated games in dynamically changing environments, motivated by application domains such as Internet ad-auctions and packet routing. We prove that, in many classes of games, if players choose their strategies in a way that guarantees low adaptive regret, then high social welfare is ensured, even under very frequent changes. In fact, in large markets learning players achieve asymptotically optimal social welfare despite high turnover. Previous work has only showed that high welfare is guaranteed for learning outcomes in static environments. Our work extends these results to more realistic settings when participation is drastically evolving over time.

cs.GT

Stability and Learning in Strategic Queuing Systems

Bounding the price of anarchy, which quantifies the damage to social welfare due to selfish behavior of the participants, has been an important area of research. In this paper, we study this phenomenon in the context of a game modeling queuing systems: routers compete for servers, where packets that do not get service will be resent at future rounds, resulting in a system where the number of packets at each round depends on the success of the routers in the previous rounds. We model this as an (infinitely) repeated game, where the system holds a state (number of packets held by each queue) that arises from the results of the previous round. We assume that routers satisfy the no-regret condition, e.g. they use learning strategies to identify the server where their packets get the best service. Classical work on repeated games makes the strong assumption that the subsequent rounds of the repeated games are independent (beyond the influence on learning from past history). The carryover effect caused by packets remaining in this system makes learning in our context result in a highly dependent random process. We analyze this random process and find that if the capacity of the servers is high enough to allow a centralized and knowledgeable scheduler to get all packets served even with double the packet arrival rate, and queues use no-regret learning algorithms, then the expected number of packets in the queues will remain bounded throughout time, assuming older packets have priority. This paper is the first to study the effect of selfish learning in a queuing system, where the learners compete for resources, but rounds are not all independent: the number of packets to be routed at each round depends on the success of the routers in the previous rounds.

cs.GT

Feedback graph regret bounds for Thompson Sampling and UCB

We study the stochastic multi-armed bandit problem with the graph-based feedback structure introduced by Mannor and Shamir. We analyze the performance of the two most prominent stochastic bandit algorithms, Thompson Sampling and Upper Confidence Bound (UCB), in the graph-based feedback setting. We show that these algorithms achieve regret guarantees that combine the graph structure and the gaps between the means of the arm distributions. Surprisingly this holds despite the fact that these algorithms do not explicitly use the graph structure to select arms; they observe the additional feedback but do not explore based on it. Towards this result we introduce a "layering technique" highlighting the commonalities in the two algorithms.

cs.LG

Learning and Trust in Auction Markets

In this paper, we study behavior of bidders in an experimental launch of a new advertising auction platform by Zillow, as Zillow switched from negotiated contracts to using auctions in several geographically isolated markets. A unique feature of this experiment is that the bidders in this market are real estate agents that bid on their own behalf, not using third-party intermediaries. To help bidders, Zillow also provided a recommendation tool that suggested a bid for each bidder. Our main focus in this paper is on the decisions of bidders whether or not to adopt the platform-provided bid recommendation. We observe that a significant proportion of bidders do not use the recommended bid. Using the bid history of the agents we infer their value, and compare the agents' regret with their actual bidding history with results they would have obtained following the recommendation. We find that for half of the agents not following the recommendation, the increased effort of experimenting with alternate bids results in increased regret, i.e., they get decreased net value out of the system. The proportion of agents not following the recommendation slowly declines as markets mature, but it remains large in most markets that we observe. We argue that the main reason for this phenomenon is the lack of trust in the platform-provided tool. Our work provides an empirical insight into possible design choices for auction-based online advertising platforms. While search advertising platforms (such as Google or Bing) allow bidders to submit bids on their own, many display advertising platforms (such as Facebook) optimize bids on bidders' behalf and eliminate the need for bids. Our empirical analysis shows that the latter approach is preferred for markets where bidders are individuals, who don't have access to third party tools, and who may question the fairness of platform-provided suggestions.

cs.GT

Simple and Efficient Budget Feasible Mechanisms for Monotone Submodular Valuations

We study the problem of a budget limited buyer who wants to buy a set of items, each from a different seller, to maximize her value. The budget feasible mechanism design problem aims to design a mechanism which incentivizes the sellers to truthfully report their cost, and maximizes the buyer's value while guaranteeing that the total payment does not exceed her budget. Such budget feasible mechanisms can model a buyer in a crowdsourcing market interested in recruiting a set of workers (sellers) to accomplish a task for her. This budget feasible mechanism design problem was introduced by Singer in 2010. There have been a number of improvements on the approximation guarantee of such mechanisms since then. We consider the general case where the buyer's valuation is a monotone submodular function. We offer two general frameworks for simple mechanisms, and by combining these frameworks, we significantly improve on the best known results for this problem, while also simplifying the analysis. For example, we improve the approximation guarantee for the general monotone submodular case from 7.91 to 5; and for the case of large markets (where each individual item has negligible value) from 3 to 2.58. More generally, given an $r$ approximation algorithm for the optimization problem (ignoring incentives), our mechanism is a $r+1$ approximation mechanism for large markets, an improvement from $2r^2$. We also provide a similar parameterized mechanism without the large market assumption, where we achieve a $4r+1$ approximation guarantee.

cs.GT

Computing Equilibrium in Matching Markets

Market equilibria of matching markets offer an intuitive and fair solution for matching problems without money with agents who have preferences over the items. Such a matching market can be viewed as a variation of Fisher market, albeit with rather peculiar preferences of agents. These preferences can be described by piece-wise linear concave (PLC) functions, which however, are not separable (due to each agent only asking for one item), are not monotone, and do not satisfy the gross substitute property-- increase in price of an item can result in increased demand for the item. Devanur and Kannan in FOCS 08 showed that market clearing prices can be found in polynomial time in markets with fixed number of items and general PLC preferences. They also consider Fischer markets with fixed number of agents (instead of fixed number of items), and give a polynomial time algorithm for this case if preferences are separable functions of the items, in addition to being PLC functions. Our main result is a polynomial time algorithm for finding market clearing prices in matching markets with fixed number of different agent preferences, despite that the utility corresponding to matching markets is not separable. We also give a simpler algorithm for the case of matching markets with fixed number of different items.

cs.GT