Meeting of COST Action IC1250 on Computational Social Choice

2-4 November 2015

Istanbul Bilgi University, İstanbul, Turkey

Programme Schedule & Speakers

MONDAY, November 2

12:45 Shuttle - Marmara Pera Hotel to Santral campus

13:00-14:00 Registration

14:00-14:15 Welcome

14:15-15:00 Remzi Sanver (Paris-Dauphine and Istanbul Bilgi)

Non-Manipulable Collective Choice with Acceptable and Unacceptable Alternatives

We consider manipulation of collective decision making rules in a preference-approval framework where voters not only rank candidates but also evaluate them as "acceptable" or "unacceptable". In this richer informational setting, we adopt a new notion of manipulability where incentives of manipulation exist if and only if a voter can ensure to replace an outcome which he finds unacceptable with an acceptable one. Under this definition, we investigate the possibility of efficient, anonymous and non-manipulable rules.

15:00-15:45 Matías Núñez (Paris-Dauphine and CNRS) [Slides]

Incentives to Foster Consensual Agreements

We focus on the economic design of fair and transparent decision rules for committees deciding over unidimensional matters. We propose the following rule: each member can vote for as many alternatives as she likes (approval voting) and the implemented policy coincides with the mean of the vote distribution (averaging process). Assuming that preferences are single-peaked, we show that this voting game has an essentially unique equilibrium in pure strategies for every possible distribution of the voters' ideal policies. We furthermore demonstrate that this equilibrium is minimally sincere (each voter votes for her ideal policy) and consensual (each voter votes for the implemented policy).

15:45-16:30 Coffe Break

16:30-17:15 Fuad Aleskerov (Higher School of Economics and Institute of Control Sciences, Moscow) [Slides]

Social Choice Rules and their Manipulability

Historical examples of manipulation are presented. Several classes of choice procedures are discussed as well as tie-breaking rules used in the case when the rule chooses more than one alternative. Multi-valuedness of aggregation rules are demonstrated. General notions of manipulability are introduced for single- and multi-valued cases as well as the indices of manipulability measuring the degree of manipulability of the rules. We define different levels of manipulability---weak and strong---in different so-called Cultures---Impartial and Impartial Anonymous Cultures. The general idea of computational schemes used in the analysis of the degree of manipulability is presented. Coalitional manipulability in Impartial Culture is studied. Comparison is made for the manipulability and decisiveness of the rules under study.

17:15-18:00 Timo Mennle (Zurich) [Slides]

The Pareto Frontier for Randomized Mechanisms

We study the trade-offs between strategyproofness and other desiderata, such as efficiency or fairness, that often arise in the design of random ordinal mechanisms. We use approximate strategyproofness to measure the manipulability of non-strategyproof mechanisms, and we introduce the deficit to quantify the performance of mechanisms with respect to a desideratum. When the desideratum is incompatible with strategyproofness, mechanisms that trade off manipulability and deficit optimally form the Pareto frontier. Our main contribution is a structural characterization of this Pareto frontier, and we present algorithms that exploit this structure to compute it. To illustrate its shape, we apply our results for two orthogonal desiderata, namely Plurality and Veto scoring, in a setting with 3 agents and 3 alternatives.

18:15 Shuttle - Santral campus to Marmara Pera Hotel

19:15 Shuttle - Santral Campus to Concert Hall

22:00 Shuttle - Concert Hall to Marmara Pera Hotel

TUESDAY, November 3

08:45 Shuttle - Marmara Pera Hotel to Santral campus

09:30-10:15 Gabrielle Demange (Paris School of Economics) [Slides]

Optimal Targeting Strategy in a Network under Positive Externalities

The goal of the present paper is to analyze the optimal targeting strategies of a planner (say, a monopolist) who aims to increase the aggregate activity of agents (consumers). Agents are embedded in a social network and their actions influence each other. Actions are complements, determined by their exposure to neighbors' actions through a response function. The description includes strategic settings and models of mechanical influence as in contagion or mimesis. When the response function is linear, it is optimal for the planner to target the same nodes, those with the largest index for a centrality measure, whatever the used resources. When the response function is not linear, concave or convex, new features determine the optimal targeting. Contrary to the linear case, the targeted nodes may vary with the planners' resources, and be more or less dispersed. Both analytical results and simulations are presented. In particular, the paper identifies key conditions on the social influence for the optimal strategy to target all nodes and evaluates the benefit drawn from the knowledge of the precise network structure.

10:15-11:00 Karol Życzkowski (Jagiellonian University) [Slides]

Jagiellonian Compromise vs. Cambridge Compromise: On Square Root and Proportional Representations

Two major mathematical issues related to the governance in the European Union are discussed: (a) voting rules in the European Council, and (b) allocation of seats in the European Parliament. Each member state is represented in the European Council by a single representative, which takes part in weighted voting with a qualified majority. We review the theory of Penrose, according to which the voting power of any citizen of any state is equal, if the voting weights are proportional to the square root of the population of each Member State. The proposed voting system, called Jagiellonian Compromise (JagCom), is based on the Penrose law and the optimal threshold of qualified majority, for EU-28 equal to 61%. In the case of the European Parliament each Member State sends several of their representatives. They vote separately, so their votes can differ to optimally represent the points of view of their electorate. This assumption leads to the linear dependence between the population of a given state and the number of Parliament members representing this state. We review certain apportionment functions and show that the constitutional constraints are so strong that admissible functions lead to rather similar solutions. The method CamCom, proposed in 2011 during a scientific meeting in Cambridge, equivalent to a fixed base plus a term proportional to the population is also discussed.

11:00-11:45 Coffe Break

11:45-13:00 Rump Session

Lihi Dery (Ariel), Marija Slavkovik (Bergen), Panos Protopapas (Lausanne), Vincent Merlin (Caen), Natalia Meshcheryakova (Moscow), Thierry Marchant (Ghent), Ana Madevska-Bogdanova (Skopje), Edith Elkind (Oxford), László Csató (Budapest), Stéphane Airiau (Paris), Justin Kruger (Paris), and Burak Can (Maastricht).

13:00-14:30 Lunch

14:30-15:15 Markus Brill (Oxford) [Slides]

Justified Representation in Approval-Based Committee Voting

We consider approval-based committee voting, i.e. the setting where each voter approves a subset of candidates, and these votes are then used to select a fixed-size set of winners (committee). We propose a natural axiom for this setting, which we call justified representation (JR). This axiom requires that if a large enough group of voters exhibits agreement by supporting the same candidate, then at least one voter in this group has an approved candidate in the winning committee. We show that for every list of ballots it is possible to select a committee that provides JR. However, it turns out that several prominent approval-based voting rules may fail to output such a committee. In particular, while Proportional Approval Voting (PAV) always outputs a committee that provides JR, Reweighted Approval Voting (RAV), which can be seen as a tractable approximation to PAV, does not have this property. We then introduce a stronger version of the JR axiom, which we call extended justified representation (EJR), and show that PAV satisfies EJR, while other rules we consider do not; indeed, EJR can be used to characterize PAV within the class of weighted PAV rules. We also consider several other questions related to JR and EJR, including the relationship between JR/EJR and core stability, and the complexity of the associated algorithmic problems. This is joint work with Haris Aziz, Vincent Conitzer, Edith Elkind, Rupert Freeman, and Toby Walsh.

15:15-16:00 Vincent Merlin (Caen)

Evaluating the Likelihood of the Referendum Paradox for Mixed Voting Systems

A referendum paradox (Nurmi, 1999) occurs, in a two party competition, each time a party gets a majority of the seats in the parliament while it did not obtain a majority of votes nationwide. This paradox can be viewed as an instance of the Borda paradox, as the voting rules fails to select the Condorcet winner. Feix, Lepelley, Merlin and Rouet (2004), Wilson and Pritchard (2007) and Lepelley, Merlin and Rouet (2011) computed the probability of the referendum paradox under the Impatial Culture (IC) assumption and a variant of the Impartial Anonymous Culture (IAC*) assumptions when two parties compete in D equal-sized districts. These a priori models for voting are extensively described in Gerhlein (2006). The same paradox may occur for mixed electoral systems. On the top of electing D representatives in districts, the voters also elect L members of the parliament at large. Hence, the parliament is of size D + L. Blais and Massicote (2009) propose an extensive survey of all the mixed electoral systems that are used worldwide. In this paper, assuming that D representatives are elected in equal-size jurisdictions, we estimate the probability of the referendum paradox for three different mixed systems: (1) when all the L at-large seats are attributed to the party which obtained a majority of votes nationwide; (2) when the L at-large seats are apportioned according to the proportional rule; and (3) when the L seats are apportioned on the basis of the wasted votes, that is, the sums of the votes of the party candidates that were not elected in districts. We perform our estimations under the IC and IAC* hypothesis. As a corollary, we estimate the probability of the referendum paradox as a function of the ratio L/D under the three scenarios. In an electoral design perspective, we are then able to suggest which values for L/D are sufficiently large for the referendum paradox to become negligible and which mixed system is more able to drastically reduce the likelihood of the paradox.
This is joint work with Dominique Lepelley and Michel Le Breton.

16:00-16:30 Coffe Break

16:30-18:00 MC Meeting

19:15 Shuttle - Santral Campus to Dinner

20:00 Dinner

22:30 Shuttle - Dinner to Marmara Pera

WEDNESDAY, November 4

09:15 Shuttle - Marmara Pera Hotel to Santral campus

10:00-10:45 Ayça Ebru Giritligil (Istanbul Bilgi)

Anonymous and Neutral Social Choice: Existence Results on Resoluteness

We present the existence conditions of anonymous, neutral and resolute social welfare and social choice rules in a group-theoretical framework. For the case where these conditions are not met, we introduce the maximum domain of preference profiles over which such aggregation rules are well-defined. We propose a procedure to obtain anonymous and neutral resolute refinements of social choice rules. Finally, we compare and contrast this procedure with two conventional tie-breaking methods which violate anonymity or neutrality, in terms of their compatibility with (simple) monotonicity. This is joint work with Onur Doğan.

10:45-11:30 Coffe Break

11:30-12:15 Maria Polukarov (Southampton)

Strategic Voting and Candidacy

We analyse voting scenarios from a game-theoretic perspective, viewing strategic parties as players and examining possible stable outcomes of their interaction (e.g., equilibria). Specifically, we model strategic behaviours by voters and candidates as voting/candidacy games, and study the existence of stable game outcomes, as well as their reachability by natural iterative processes, such as best-response dynamics or its restricted variants. Convergence of such procedures is a highly desirable property of the game, since, from a system-wide perspective, it implies that a system has a deterministic stable state that can be reached by the agents without any centralised control and/or communication.

12:15-13:00 Reshef Meir (Technion) [Slides]

Uncertainty and Bounded-Rationality in Voting

In multi-agent interactions, voting included, each agent often faces uncertainty over the incentives and the behavior of the other agents. The traditional approach (e.g. Myerson and Weber '93) assumes that voters each maximize their expected utility w.r.t. some common prior distribution regarding others' preferences. However in most real-world scenarios voters have no way to accurately or even approximately know this distribution. Moreover, numerous psychological experiments have demonstrated that human decision makers fail even at fairly simple tasks involving probabilistic reasoning, and are prone to cognitive biases such as risk-aversion and loss-aversion. I will describe an alternative, non-probabilistic, model for representing players' uncertainty in games, inspired by artificial intelligence and bounded rationality approaches. While the model is quite general, I will demonstrate how it applies for the Plurality voting mechanism, overcoming many shortcomings of previous theories. My main result is that the behavior of bounded-rational agents boils down to a simple and natural dynamics, which is guaranteed to converge to an equilibrium. Extensive simulations show that the resulting equilibria replicate known phenomena from real-world voting. A nice property of the model is its flexibility, which allows for natural extensions to other voting rules, different number of winners, limited or structured information flow, and so on. If time permits, I will show how key components of this approach can be extracted and applied to understand strategic behavior of participants on the popular scheduling platform Doodle.
The talk is based on joint work with Omer Lev, David Parkes, Jeffrey S. Rosenschein, and James Zou.

14:00 Shuttle - Santral campus to Marmara Pera Hotel

Practical Informations

Hotel Informations

There will be a shuttle service between Taksim Pera region and venue of the workshop. The following are in Taksim Pera region that are walking distance to shuttle service.

The Marmara Pera

Please contact Halil İbrahim Pehlivan at hpehlivan@themarmarahotels.com until 28.09.2015 with the code “BIPERA” to take advantage of group prices: 120€ per night for single, 130€ per night for double including breakfast and VAT.

The Peak Hotel

Please contact sales@thepeakhotel.com.tr with the code “COSTIST“to take advantage of group prices: Eccentric Double 70€ per night and Comfort Double 80€ per night including breakfast and VAT.

Richmond Istanbul

Other Hotels in the region
Troya Hotel
Grand Hotel de Pera
Radisson Blu Pera

Venue

The workshop shall take place in Energy Museum Seminar Room at Santral Campus of Istanbul Bilgi University. See the map below.

How to reach campus

Transportation

From Ataturk airport to Campus

Public transportation: Take Havatas airport shuttle to Taksim. It departs from airport every half an hour and cost 11 TL (4 Euro). From Taksim square, taxi to campus should cost 20 TL (6 Euro). Alternatively, you can take finucular from Taksim Square to Kabatas and from Kabatas, you can take University shuttles.

Taxi From airport to campus should cost you around 60 TL (20 Euro).

Webpage for airport shuttle Havatas
Webpage for university shuttles

From Ataturk airport to Pera (Hotel Region)

Pera is in walking distance to Taksim Square. You can take airport shuttle Havatas to Taksim square. Details for Havatas are above.

Taxi from airport to Pera should also cost you around 60 TL (20 Euro).

From Pera Region to Campus

There will be shuttles between Pera and Campus before and after meeting. The exact time table will be posted later.

If you have to take public transportation, then walk to Taksim Square and take funicular to Kabatas and from Kabatas, take the university shuttles.

Local organizer

You can concact with Murat Ali Cengelci at murat.cengelci@bilgi.edu.tr

Venue

Eski Silahtarağa Elektrik Santralı Kazım Karabekir Cad. No: 2/13 34060 Eyüp İstanbul
444 0 428