We present the design of a banner advertising auction which is considerably more expressive than current designs. We describe a general model of expressive ad contracts/bidding and an allocation model that can be executed in real time through the assignment of fractions of relevant ad channels to specific advertiser contracts. The uncertainty in channel supply and demand is addressed by the formulation of a stochastic combinatorial optimization problem for channel allocation that is rerun periodically. We solve this in two different ways: fast deterministic optimization with respect to expectations; and a novel online sample-based stochastic optimization method— that can be applied to continuous decision spaces—which exploits the deterministic optimization as a black box. Experiments demonstrate the importance of expressive bidding and the value of stochastic optimization. 1

When online ads are shown together, they compete for user attention and conversions, imposing negative externalities on each other. We study the simplest form that an advertiser’s valuation can have in the presence of such externalities – an advertiser’s value depends on whether her ad is shown exclusively, or along with other ads. A mechanism such as the GSP auction, which always displays a full slate of ads, can be arbitrarily inefficient given such two-dimensional valuations; further, expanding the outcome space need not be enough– when the advertiser pays for an impression or a click, but derives value from a conversion, the private value of the advertiser is affected by the externality as well, leading to the need for a more expressive bidding language. In this paper, we study the design of expressive mechanisms which are extensions of the GSP auction currently used in practice. Our mechanisms show either a single ad exclusively, or multiple ads simultaneously, with the property that the allocation and pricing are identical to GSP when multiple ads are shown. We investigate the equilibria of these mechanisms, and show that the revenue when multiple ads are shown dominates the VCG revenue. In equilibria where a single ad is shown, the revenue can actually be lower than that of VCG; however, this loss is bounded within a factor of two. The increased efficiency from using the more expressive mechanism can come, unfortunately, with a loss in revenue with respect to the existing GSP mechanism in some cases. We design a mechanism which has a one-dimensional bidding language, while still allowing two types of outcomes, for situations where the private values are approximately one-dimensional. We show that this mechanism, while retaining the revenue and efficiency properties of the previous mechanisms, also revenue dominates the existing GSP mechanism. Part of this work was done while Amin Sayedi was visiting

We present a three-week user study in which we tracked the locations of 27 subjects and asked them to rate when, where, and with whom they would have been comfortable sharing their locations. The results of analysis conducted on over 7,500 hours of data suggest that the user population represented by our subjects has rich location-privacy preferences, with a number of critical dimensions, including time of day, day of week, and location. We describe a methodology for quantifying the effects, in terms of accuracy and amount of information shared, of privacy-setting types with differing levels of complexity (e.g., setting types that allow users to specify location- and/or time-based rules). Using the detailed preferences we collected, we identify the best possible policy (or collection of rules granting access to one’s location) for each subject and privacy-setting type. We measure the accuracy with which the resulting policies are able to capture our subjects ’ privacy preferences under a variety of assumptions about the sensitivity of the information and user-burden tolerance. One practical implication of our results is that today’s location sharing applications may have failed to gain much traction due to their limited privacy settings, as they appear to be ineffective at capturing the preferences revealed by our study.

A key trend in (electronic) commerce is a demand for higher levels of expressiveness in the mechanisms that mediate interactions. Recent results indicate the increase in expressiveness tends to lead to increased efficiency. Online advertisement (ad) auctions account for tens of billions of dollars in revenue annually and are some of the fastest growing mechanisms on the Internet. However, the most frequent variant of these mechanisms does not allow bidders (agents) to offer a separate bid for each ad position, and is thus inexpressive on a fundamental level. In this paper we attempt to characterize the cost of this inexpressiveness. We adapt a theoretical framework to show that the commonly used generalized second price (GSP) mechanism is arbitrarily inefficient for some distributions over agent preferences. We then describe a search technique that computes an upper bound on the expected efficiency of the GSP mechanism for a given distribution over agent preferences. We report the results of running our search technique on synthetic preference distributions. Our results demonstrate that the cost of inexpressiveness is most severe when agents have diverse preferences and relatively low profit margins. Our results also show that designating one or more positions as “premium ” and soliciting an extra bid for these positions eliminates almost all of the inefficiency. 1.

Information and Communication Technologies Institute. The authors would also like to thank Lucian Cesca,

Mechanisms (especially on the Internet) have begun allowing people or organizations to express richer preferences in order to provide for greater levels of overall satisfaction. In this paper, we develop an operational methodology for quantifying the expected gains in economic efficiency associated with different forms of expressiveness. We begin by proving that the sponsored search mechanism (GSP) used by Google, Yahoo!, MSN, etc. can be arbitrarily inefficient. We then experimentally compare its efficiency to a slightly more expressive variant (PGSP), which solicits an extra bid for a premium class of positions. We generate random preference distributions based on published industry knowledge. We determine ideal strategies for the agents using a custom tree search technique, and we also benchmark using straightforward heuristic bidding strategies. The GSP’s efficiency loss is greatest in the practical case where some advertisers (“brand advertisers”) prefer top positions while others (“value advertisers”) prefer middle positions, and that loss can be dramatic. It is also worst when agents have small profit margins. While the PGSP is only slightly more expressive (and thus not much more cumbersome), it removes almost all of the efficiency loss in all of the settings we study. 1

Cake cutting is a playful name for the problem of fairly dividing a heterogeneous divisible good among a set of agents. The agent valuations for different pieces of cake are typically assumed to be additive. However, in certain practical settings this assumption is invalid because agents may not have positive value for arbitrarily small “crumbs” of cake. In this paper, we propose a new, more expressive model of agent valuations that captures this feature. We present an approximately proportional algorithm for any number of agents that have such expressive valuations. The algorithm is optimal in the sense that no other algorithm can guarantee a greater worst-case degree of proportionality. We also design an optimal approximately proportional and fully envy-free algorithm for two agents. 1

Everything should be made as simple as possible, but no simpler.

A recent trend (especially in electronic commerce) is higher levels of expressiveness in the mechanisms that mediate interactions such as auctions, exchanges, catalog offers, voting systems, matching of peers, and so on. Participants can express their preferences in drastically greater detail than ever before. In many cases this trend is fueled by modern algorithms for winner determination that can handle the richer inputs. But is more expressiveness always a good thing? What forms of expressiveness should be offered? In this talk I will first report on our experience from over $40 billion of combinatorial multi-attribute sourcing auctions. Then, I will present recent theory that ties the expressiveness of a mechanism to an upper bound on efficiency in a domain-independent way in private-information settings. Time permitting, I will also discuss theory and experiments on applying expressiveness to ad auctions, such as sponsored search and real-time banner ad auctions with temporal span and complex preferences. 1

Mechanisms are present everywhere in both business and social contexts. They govern the interactions people have with businesses, governments, and each other. One emerging trend over the past decade is a demand for higher levels of expressiveness in mechanisms that mediate interactions such as the allocation of resources, matching of peers, and elicitation of opinions. This trend has already manifested itself in combinatorial auctions and generalizations thereof. It is also reflected in the richness of preference expression offered by businesses as diverse as matchmaking sites, sites like Amazon and Netflix, and services like Google’s AdSense. In Web 2.0 parlance, this demand for increasingly diverse offerings is called the Long Tail. A driving force behind this trend is that greater expressiveness begets better matches, or greater efficiency of the outcomes. Yet, expressiveness does not come for free; it burdens users to specify more preference information. Today’s mechanisms have relied on empirical tweaking to determine how to deal with this and related tradeoffs. In this thesis, we propose to establish the foundations of expressiveness in mechanisms and its relationship to their efficiency,