BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//6.3//EN
TZID:Asia/Jerusalem
X-WR-TIMEZONE:Asia/Jerusalem
BEGIN:VEVENT
UID:17@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20201028T113000
DTEND;TZID=Asia/Jerusalem:20201028T123000
DTSTAMP:20201109T060748Z
URL:https://dds.technion.ac.il/ar/iemevents/when-both-players-choose-domin
ated-actions-underweighting-rare-events-in-2-player-repeated-games/
SUMMARY:When both players choose dominated actions: Underweighting rare eve
nts in 2-player repeated games [ \n Game Theory Seminar\n Seminars
\n \n ]
DESCRIPTION:By: Ori Plonsky\n Advisors: \n Where: Zoom From:\nTechnion\nA
bstract:\n\nBehavioral decision studies of individuals who make repeated d
ecisions with feedback reveal robust evidence that people behave as if the
y believe “it won’t happen to me”\, a phenomenon coined underweighti
ng of rare events. We experimentally show that the tendency to underweight
rare events persists in 2-person repeated games with stochastic payoffs\,
and that other agents can learn to exploit it. In a simultaneous-move asy
mmetric 2x2 game\, most row players consistently choose a stochastically d
ominated action that provides a better payoff most of the time but on aver
age leads to a large loss over the equilibrium prediction\, behavior consi
stent with underweighting of the rare event. In response\, most column pla
yers learn to exploit this bias by choosing a strictly dominated action th
at is worse for the row players on average but better most of the time. Th
at is\, due to underweighting of rare events by the row players\, most dya
ds converge to a profile of two dominated strategies. A second study rules
out different explanations like boredom\, altruism or risk seeking.\n\nJo
int work with Yefim Roth (University of Haifa)\n\nLink to Seminar Zoom
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:23@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20201111T113000
DTEND;TZID=Asia/Jerusalem:20201111T123000
DTSTAMP:20201109T060933Z
URL:https://dds.technion.ac.il/ar/iemevents/an-algorithmic-framework-for-a
pproximating-maximin-share-allocation-of-chores/
SUMMARY:An Algorithmic Framework for Approximating Maximin Share Allocation
of Chores [ \n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Xin Huang\n Advisors: \n Where: ZOOM From:\nTechnion\nAbst
ract:\n\nIn this paper\, we consider the problem of how to fairly dividing
m indivisible chores among n agents. The fairness measure we considered h
ere is the maximin share. The previous best known result is that there alw
ays exists a 4/3 approximation maximin share allocation. With our algorit
hm\, we can always find a 11/9 approximation maximin share allocation for
any instance. We also discuss how to improve the efficiency of the algorit
hm and its connection to the job scheduling problem.\n\nSeminar Zoom Link
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:28@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20201125T113000
DTEND;TZID=Asia/Jerusalem:20201125T123000
DTSTAMP:20201118T173908Z
URL:https://dds.technion.ac.il/ar/iemevents/fair-cake-division/
SUMMARY:Fair Cake Division [ \n Game Theory Seminar\n Seminars\n
\n ]
DESCRIPTION:By: Nidhi Rathi \n Advisors: \n Where: ZOOM From:\nthe Indian
Institute of Science\nAbstract:\n\nThe classic cake-cutting problem provid
es a model for addressing fair allocation of a divisible resource (metapho
rically\, the cake) among agents with distinct preferences. Envy-free (fai
r) cake divisions with contiguous pieces are known to exist under mild con
ditions\, but it is computationally hard to find them. In this talk\, I wi
ll present two of my recent results which complements these existential (a
nd non-constructive) guarantees by developing polynomial-time approximatio
n algorithms and by identifying computationally tractable instances for fa
ir cake division. First\, I will discuss an efficient algorithm for findin
g a cake division whose envy is multiplicatively bounded by 1/3. Moving fo
rward\, I will present a result that develops efficient cake-cutting algor
ithms to find envy-free divisions for a broad class of valuations (that sa
tisfies the monotone likelihood ratios property). In particular\, our algo
rithmic result holds when the agents' valuations are induced by linear tra
nslations of any log-concave function\, such as Gaussian\, exponential\, l
inear\, or binomial.\n\n \;\n\nJoint work with Siddharth Barman\, Eshw
ar Ram Arunchaleswaran and Rachitesh Kumar\n\n \;\n\nZoom Link
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:31@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20201209T113000
DTEND;TZID=Asia/Jerusalem:20201209T113000
DTSTAMP:20201203T070657Z
URL:https://dds.technion.ac.il/ar/iemevents/incomplete-information-vcg-con
tracts-for-common-agency/
SUMMARY:Incomplete Information VCG Contracts for Common Agency [ \n Gam
e Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Tal Alon \n Advisors: \n Where: Zoom From:\nTechnion\nAbst
ract:\nWe study the social efficiency of contracts as economic mechanisms
when multiple principals simultaneously manage a common agent. We consider
an incomplete-information setting: The agent chooses an unobservable acti
on that induces both a cost for the agent and an expected value for each p
rincipal. The sum of these terms is referred to as the resulting social we
lfare. The agent’s choice is incentivized by payment schemes (“contrac
ts”) set forth by the principals\, who have different private values for
different stochastic outcomes of the agent’s choice. We enforce two sta
ndard properties of contracts: individual rationality a.k.a. IR (for the p
rincipals)\, and limited liability a.k.a. LL (for the agent).\n\nJoint wor
k with Ron Lavi\, Elisheva Shamash\, and Inbal Talgam-Cohen\n\nZoom Link
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:39@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20201223T113000
DTEND;TZID=Asia/Jerusalem:20201223T123000
DTSTAMP:20201220T115150Z
URL:https://dds.technion.ac.il/ar/iemevents/settling-the-complexity-of-nas
h-equilibrium-in-congestion-games/
SUMMARY:A Market-Inspired Bidding Scheme for Peer Review Paper Assignment [
\n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Reshef Meir \n Advisors: \n Where: Zoom From:\nTechnion\nA
bstract:\n\nWe propose a market-inspired bidding scheme for the assignment
of paper reviews in large academic conferences. We provide an analysis of
the incentives of reviewers during the bidding phase\, when reviewers hav
e both private costs and some information about the demand for each paper\
; and their goal is to obtain the best possible k papers for a predetermin
ed k.\n\nWe show that by assigning `budgets' to reviewers and a `price' fo
r every paper that is (roughly) proportional to its demand\, the best resp
onse of a reviewer is to bid sincerely\, i.e.\, on her most favorite pape
rs\, and match the budget even when it is not enforced. This game-theore
tic analysis is based on a simple extension of the Trading Post mechanism.
\nWe show via extensive simulations on bidding data from real conferences\
, that our bidding scheme would substantially improve both the bid distr
ibution and the resulting assignment.\n\n \;\n\njoint work with Jerome
Lang\, Lulien Lesca\, Nick Mattei\, and Natan Kaminsky\n\nZoom Link\n\n&n
bsp\;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:40@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20210106T113000
DTEND;TZID=Asia/Jerusalem:20210106T123000
DTSTAMP:20201230T172116Z
URL:https://dds.technion.ac.il/ar/iemevents/settling-the-complexity-of-nas
h-equilibrium-in-congestion-games-2/
SUMMARY:Settling the complexity of Nash equilibrium in congestion games [ \
n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Yakov Babichenko \n Advisors: \n Where: Zoom From:\nTechni
on\nAbstract:\n\nWe consider (i) the problem of finding a (possibly mixed)
Nash equilibrium in congestion game\, and (ii) the problem of finding an
(exponential precision) fixed point of the gradient descent dynamics of a
smooth function f ∶ [0\, 1]^n → R. We prove that these problems are eq
uivalent. Our result holds for various explicit descriptions of f\, rangin
g from (almost general) arithmetic circuits\, to degree-5 polynomials. By
a very recent result of [FGHS20]\, this implies that these problems are PP
AD ∩ PLS-complete. As a corollary\, we also obtain the following equival
ence of complexity classes:\n\nCCLS = PPAD∩PLS\n\nZoom https://technion.
zoom.us/j/95592625013
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:51@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20210120T113000
DTEND;TZID=Asia/Jerusalem:20210120T123000
DTSTAMP:20210113T082053Z
URL:https://dds.technion.ac.il/ar/iemevents/prophet-and-secretary-online-a
lgorithms-for-matching-in-general-graphs/
SUMMARY:Prophet and Secretary Online Algorithms for Matching in General Gra
phs [ \n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Michal Feldman \n Advisors: \n Where: ZOOM From:\nTel-Avi
v University\nAbstract:\n\nProphet and Secretary Online Algorithms for Mat
ching in General Graphs\n\nA common tension in market scenarios is choosin
g the right timing to commit to a decision. This tension is captured by th
e mathematical literature of optimal stopping theory\, within two models t
hat have become to be known as the secretary problem and the prophet inequ
ality. In these models\, a sequence of originally-unknown values arrive on
e by one. Upon arrival\, the online algorithm observes the value and shoul
d decide whether or not to accept it. In secretary settings\, the value se
quence is arbitrary\, but the values arrive in a uniformly random order. I
n prophet settings\, every value is drawn from a known probability distrib
ution\, but the arrival order is arbitrary.\n\nIn this talk I will review
the basic settings of secretary and prophet\, as well as previous extensio
ns to matching in bipartite graphs with 1-sided vertex arrival. I will the
n present our recent work\, which studies online algorithms (in both secre
tary and prophet settings) for matching in *general* graphs\, under both v
ertex- and edge-arrival models. We provide tight competitive ratios for bo
th secretary and prophet matching scenarios under vertex arrival. Under ed
ge arrival\, we provide competitive ratios that improve upon the state of
the art.\n\nBased on the following joint work with Tomer Ezra\, Nick Gravi
n\, and Zhihao Tang:\n\nhttps://arxiv.org/abs/2002.09807\n\nhttps://arxiv.
org/abs/2011.01559\n\nBe rational and attend\n\nZoom Link\n\nhttps://techn
ion.zoom.us/j/95592625013
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:74@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20210519T113000
DTEND;TZID=Asia/Jerusalem:20210519T123000
DTSTAMP:20210512T130803Z
URL:https://dds.technion.ac.il/ar/iemevents/collective-information-acquisi
tion/
SUMMARY:Collective Information Acquisition [ \n Game Theory Seminar\n
Seminars\n \n ]
DESCRIPTION:By: Ran Wilat \n Advisors: \n Where: Bloomfield 424 From:\nBe
n-Gurion University\nAbstract:\n\nWe consider the problem faced by a group
of players who bargain over what public signal to acquire before deciding
on a collective action. The players differ in their privately known state
-dependent payoffs from taking the action\, and therefore differ also in t
heir willingness to pay for the public signal. We take a mechanism design
approach to characterize the frontier of outcomes achievable via bargainin
g over information. We identify novel distortions in the optimal informati
on structure that arise from the information asymmetry and because\, after
the signal is realized\, the outcome is determined in equilibrium of a su
bsequent voting game.\n\n(joint with K. Eliaz)
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:76@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20210602T113000
DTEND;TZID=Asia/Jerusalem:20210602T123000
DTSTAMP:20210530T052855Z
URL:https://dds.technion.ac.il/ar/iemevents/informative-tests-in-signaling
-environments/
SUMMARY:Informative Tests in Signaling Environments [ \n Game Theory Se
minar\n Seminars\n \n ]
DESCRIPTION:By: Ran Weksler\n Advisors: \n Where: Bloomfield 424 From:\nU
niversity of Haifa\nAbstract:\n\nWe study a receiver’s learning problem
of choosing an informative test in a signaling environment. Each test that
the receiver chooses induces a signaling subgame. Thus\, in addition to t
he direct effect of the chosen test on the information that the receiver o
btains\, it also has an indirect effect on the receiver’s information th
rough the sender’s signaling strategy. We analyze how signaling consider
ations affect the receiver’s preference relation over tests. Specificall
y\, we find that the receiver’s preference relation does not comply with
Blackwell’s order. Our findings may help shed light on phenomena such a
s grade inflation and information coarsening.\n\n \;\n\n(joint with B.
Zick)
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:81@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20210616T113000
DTEND;TZID=Asia/Jerusalem:20210616T123000
DTSTAMP:20210609T064041Z
URL:https://dds.technion.ac.il/ar/iemevents/barter-exchange-with-length-se
nsitive-utility-functions/
SUMMARY:Barter Exchange with Length Sensitive Utility Functions [ \n Ga
me Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Matan-el Shpiro \n Advisors: \n Where: Bloomfield 424 Fro
m:\nTechnion\nAbstract:\n\nConsider a \\emph{barter exchange} problem over
a finite set of agents\, where each agent owns an item and is also associ
ated with a (privately known) \\emph{wish list} of items belonging to the
other agents. An outcome of the problem is a (re)allocation of the items t
o the agents such that each agent either keeps her own item or receives an
item from her (reported) wish list\, subject to the constraint that the l
ength of the trading cycles induced by the allocation is up-bounded by a p
respecified \\emph{length bound} $k$. The utility of an agent from an allo
cation is $0$ if she keeps her own item\; and it is $\\lambda(\\ell)$ if s
he receives an item from her (true) wish list and participates in a tradin
g cycle of length $2 \\leq \\ell \\leq k$\, where $\\lambda$ is a prespeci
fied monotonically non-increasing \\emph{length function}. (The agent incu
rs a large dis-utility if she receives an item that is neither hers nor be
longs to her wish list.)\n\nIn this talk\, we investigate the aforemention
ed barter exchange problem from the perspective of mechanism design withou
t money\, aiming for truthful (and individually rational) mechanisms whose
objective is to maximize the \\emph{social welfare}. As the construction
of a social welfare maximizing allocation is computationally intractable f
or length bounds $k \\geq 3$\, we focus on (computationally efficient) tru
thful mechanisms that approximate the optimal social welfare. In particula
r\, we establish upper and lower bounds on the guaranteed approximation ra
tio\, expressed in terms of the length bound $k$ and the length function $
\\lambda$. Our main technical contribution is an algorithmic tool that can
be viewed as a truthful version of the \\emph{local search} paradigm.\n\n
(joint with Y. Emek)
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:104@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20210707T113000
DTEND;TZID=Asia/Jerusalem:20210707T113000
DTSTAMP:20210630T095052Z
URL:https://dds.technion.ac.il/ar/iemevents/learner-private-sequential-lea
rning-and-convex-optimization/
SUMMARY:Learner-Private Sequential Learning and Convex Optimization [ \n
Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Kuang Xu \n Advisors: \n Where: Bloomfield 424 From:\nSta
nford University\nAbstract:\n\nThe increasing ubiquity of large-scale infr
astructures for surveillance and data analysis has made understanding the
impact of privacy a pressing priority. We propose a framework for studying
a fundamental query complexity versus privacy tradeoff in sequential lea
rning. The central question is: how can one perform learning in such a man
ner that makes sure that an overseeing adversary cannot obtain the learned
model by observing the queries. We will examine two formulations of the
model (deterministic and Bayesian)\, and in both cases establish matching
upper and lower bounds on the optimal query complexity\, for a given level
of privacy guarantees. The analysis will exploit on some essential combi
natorial as well as information theoretic structures of the problem\, whic
h we will also discuss.\n\nPapers:\nhttps://arxiv.org/pdf/1805.02136.pdf\n
\nhttps://arxiv.org/abs/1911.06903\n\nhttps://arxiv.org/abs/1909.09836\n\n
https://arxiv.org/pdf/2102.11976.pdf\n\n \;\n\nBio: Kuang Xu is an
Associate Professor of Operations\, Information and Technology at Stanford
Graduate School of Business\, and Associate Professor by courtesy with th
e Electrical Engineering Department\, Stanford University. Born in Suzhou
\, China\, he received the B.S. degree in Electrical Engineering (2009) fr
om the University of Illinois at Urbana-Champaign\, and the Ph.D. degree i
n Electrical Engineering and Computer Science (2014) from the Massachusett
s Institute of Technology. His research focuses on understanding fundamen
tal properties and design principles of large-scale stochastic systems usi
ng tools from probability theory and optimization\, with applications in q
ueueing networks\, privacy and machine learning. He is a recipient of the
First Place in the INFORMS George E. Nicholson Student Paper Competition
(2011)\, the Best Paper Award\, as well as the Kenneth C. Sevcik Outstandi
ng Student Paper Award at ACM SIGMETRICS (2013)\, and the ACM SIGMETRICS
Rising Star Research Award (2020). He currently serves as an Associate Edi
tor for Operations Research.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:108@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20210715T103000
DTEND;TZID=Asia/Jerusalem:20210715T113000
DTSTAMP:20210711T061737Z
URL:https://dds.technion.ac.il/ar/iemevents/equitable-voting-rules/
SUMMARY:Equitable voting rules [ \n Game Theory Seminar\n Seminars\
n \n ]
DESCRIPTION:By: Omer Tamuz \n Advisors: \n Where: Bloomfield 424 and ZOOM
From:\nCaltech\nAbstract:\nMay’s Theorem (1952)\, a celebrated result in
social choice\, provides the foundation for majority rule. May’s crucia
l assumption of symmetry\, often thought of as a procedural equity require
ment\, is violated by many choice procedures that grant voters identical r
oles. We show that a weakening of May’s symmetry assumption allows for a
far richer set of rules that still treat voters equally. We show that suc
h rules can have minimal winning coalitions comprising a vanishing fractio
n of the population\, but not less than the square root of the population
size. Methodologically\, we introduce techniques from group theory and ill
ustrate their usefulness for the analysis of social choice questions.\n(wi
th Laurent Bartholdi\, Wade Hann-Caruthers\, Maya Josyula\, and Leeat Yari
v)\n\nZoom Link\n\nhttps://technion.zoom.us/j/94614628255
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:117@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211012T143000
DTEND;TZID=Asia/Jerusalem:20211012T153000
DTSTAMP:20211014T055338Z
URL:https://dds.technion.ac.il/ar/iemevents/communicating-with-anecdotes/
SUMMARY:Communicating with Anecdotes [ \n Game Theory Seminar\n Sem
inars\n \n ]
DESCRIPTION:By: Nicole Immorlica \n Advisors: \n Where: ZOOM From:\nMicros
oft Research\nAbstract:\n\nWe study a model of social learning and communi
cation using hard anecdotal evidence. There are two Bayesian agents (a sen
der and a receiver) who wish to communicate. The receiver must take an act
ion whose payoff depends on their personal preferences and an unknown stat
e of the world. The sender has access to a collection of samples correlate
d with the state of the world\, which we think of as specific anecdotes or
pieces of evidence\, and can send exactly one of these samples to the rec
eiver in order to influence her choice of action. Importantly\, the sender
's personal preferences may differ from the receiver's\, which affects the
sender's strategic choice of which anecdote to send. We show that if the
sender has commitment power\, then they will choose an unbiased and maxima
lly informative communication scheme\, no matter the difference in prefere
nces. Without observability\, however\, even a small difference in prefere
nces can lead to a significant bias in the choice of anecdote\, which the
receiver must then account for. This can significantly reduce the informat
iveness of the signal\, leading to substantial utility loss for both sides
. One implication is informational homophily: a receiver can rationally pr
efer to obtain information from a poorly-informed sender with aligned pref
erences\, rather than a knowledgeable expert whose preferences may differ
from her own.\n\nJoint work with Nika Haghtalab\, Brendan Lucier\, Markus
Mobius\, and Divyarthi Mohan.\n\n \;\n\nZoom Link\n\nhttps://tau-ac-il
.zoom.us/j/84735509601
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:126@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211103T113000
DTEND;TZID=Asia/Jerusalem:20211103T123000
DTSTAMP:20211028T141830Z
URL:https://dds.technion.ac.il/ar/iemevents/sampling-dynamics-and-stable-m
ixing-in-hawk-dove-games/
SUMMARY:Sampling Dynamics and Stable Mixing in Hawk–Dove Games [ \n G
ame Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Srinivas Arigapudi\n Advisors: \n Where: Bloomfield 424 F
rom:\nTechnion\nAbstract:\n\nThe hawk–dove game admits two types of equi
libria: an asymmetric pure equilibrium\nin which players in one population
play “hawk” and players in the other population play “dove\,” and
\na symmetric mixed equilibrium. The existing literature on dynamic evolut
ionary models shows that\npopulations will converge to playing one of the
asymmetric pure equilibria from any initial state.\nBy contrast\, we show
that plausible sampling dynamics\, in which agents occasionally revise the
ir actions\nby observing either opponents’ behavior or payoffs in a few
past interactions\, can\ninduce the opposite result: global convergence to
a symmetric mixed equilibrium.\n(with Yuval Heller and Amnon Schreiber)
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:129@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211110T113000
DTEND;TZID=Asia/Jerusalem:20211110T123000
DTSTAMP:20211104T082354Z
URL:https://dds.technion.ac.il/ar/iemevents/auctions-with-multi-member-bid
ders/
SUMMARY:Auctions with multi-member bidders [ \n Game Theory Seminar\n
Seminars\n \n ]
DESCRIPTION:By: Shiran Rachmilevitch \n Advisors: \n Where: Bloomfield 424
From:\nUniversity of Haifa\nAbstract:\n\nA second-price auction is consi
dered\, in which one of the bidders is a team consisting of several indivi
duals. For these individuals\, the auctioned item is a public good\, in th
e sense that either all of them win it together or all lose. They need to
agree on a bid\, and on splitting the payment to the auctioneer if they wi
n the item. If the competition that the team faces is fierce enough\, then
the model has a unique equilibrium. The equilibrium\, which is symmetric\
, generalizes the weak dominance equilibrium of the ordinary second-price
auction and satisfies intuitive comparative static properties with respect
to the team’s size. A certain modification of the model gives rise to a
symmetric equilibria\, in which only one team member participates in the a
uction and everybody else free ride.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:133@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211117T113000
DTEND;TZID=Asia/Jerusalem:20211117T123000
DTSTAMP:20211114T070011Z
URL:https://dds.technion.ac.il/ar/iemevents/strategic-classification-and-t
he-allegory-of-the-cave/
SUMMARY:Strategic Classification and the Allegory of the Cave [ \n Game
Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Assistant Professor Nir Rosenfeld \n Advisors: \n Where: B
loomfield 424 From:\nTechnion\nAbstract:\n\nAcross a multitude of domains
and applications\, machine learning has become widespread as a tool for i
nforming decisions about humans and for humans. Nonetheless\, tools used i
n practice are rarely designed to account for the ways in which humans res
pond to learned models and their predictions. The goal of this talk is to
initiate discussion regarding the part game theory can play in the develop
ment of behaviorally-aware learning frameworks.\nAs a starting point\, I w
ill present the problem of strategic classification\, in which users strat
egically modify their features (at a cost) to obtain favorable predictive
outcomes from a classifier. I will then describe some of our work in this
field\, both recent and current\, using these to demonstrate how even mild
forms of strategic behavior can dramatically transform the learning probl
em. Finally\, I will argue that while the original problem formulation is
somewhat narrow\, there is untapped potential in strategic classification
as a basis on which more elaborate forms of economic modeling can be appli
ed.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:135@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211124T113000
DTEND;TZID=Asia/Jerusalem:20211124T123000
DTSTAMP:20211124T061619Z
URL:https://dds.technion.ac.il/ar/iemevents/the-gatekeepers-effect/
SUMMARY:The Gatekeeper's Effect [ \n Game Theory Seminar\n Seminars
\n \n ]
DESCRIPTION:By: Dr. Moran Koren \n Advisors: \n Where: Bloomfield 424 and
ZOOM From:\nTel Aviv University\nAbstract:\n\nMany selection processes co
ntain a ``gatekeeper". The gatekeeper's goal is to examine an applicant's
suitability for a position before both parties incur substantial costs.
Intuitively\, a gatekeeper should reduce selection costs by sifting unli
kely applicants. However\, as we show\, this is not always the case sinc
e the gatekeeper's introduction inadvertently interferes with the candidat
e's self-selection. We study the conditions under which a gatekeeper im
proves the system's efficiency and those under which it induces inefficien
cy. Additionally\, we show that selection correctness can\, at times\, b
e improved by allowing for strategic gatekeeping.\n\n \;\n\n \;\n\
nZoom Link\n\nhttps://technion.zoom.us/j/92624364096
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:143@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211208T113000
DTEND;TZID=Asia/Jerusalem:20211208T123000
DTSTAMP:20211206T063154Z
URL:https://dds.technion.ac.il/ar/iemevents/markovian-persuasion/
SUMMARY:Markovian Persuasion [ \n Game Theory Seminar\n Seminars\n
\n ]
DESCRIPTION:By: Dimitry Shaiderman \n Advisors: \n Where: Zoom From:\nTel
Aviv University\nAbstract:\n\nIn the classical Bayesian persuasion model a
n informed player and an uninformed one engage in a static interaction. T
he informed player\, the sender\, knows the state of nature\, while the
uninformed one\, the receiver\, does not. The informed player partially sh
ares his private information with the receiver and the latter then\, base
d on her belief about the state\, takes an action. This action determines
\, together with the state of nature\, the utility of both players. We co
nsider a dynamic Bayesian persuasion situation where the state of nature e
volves according to a Markovian law. In this repeated persuasion model an
optimal disclosure strategy of the sender should\, at any period\, balanc
e between getting high stage payoff and future implications on the receiv
ers' beliefs. We discuss optimal strategies under different discount fact
ors and characterize when the asymptotic value achieves the maximal value
possible.\n\nJoint work with Ehud Lehrer\n\nlink to the paper\n\n \;
\n\nLink to Zoom\nhttps://technion.zoom.us/j/92624364096
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:144@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211215T113000
DTEND;TZID=Asia/Jerusalem:20211215T123000
DTSTAMP:20211209T102211Z
URL:https://dds.technion.ac.il/ar/iemevents/stochastic-games-with-shift-in
variant-payoffs/
SUMMARY:Stochastic Games with Shift-Invariant Payoffs [ \n Game Theory
Seminar\n Seminars\n \n ]
DESCRIPTION:By: Prof. Eilon Solan\n Advisors: \n Where: Bloomfield 424 a
nd ZOOM From:\nTel Aviv University\nAbstract:\n\nI will describe recent re
sults on the existence and characterization of equilibrium payoffs in mult
iplayer repeated games with tail-measurable payoffs\, and then discuss an
extension to stochastic games: two-player nonzero-sum stochastic games wit
h shift-invariant payoffs.\n\nJoint work with Janos Flesch\n\nZoom Link\n\
nhttps://technion.zoom.us/j/92624364096
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:147@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211222T113000
DTEND;TZID=Asia/Jerusalem:20211222T123000
DTSTAMP:20211216T130254Z
URL:https://dds.technion.ac.il/ar/iemevents/herd-design/
SUMMARY:Herd Design [ \n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Ronen Gradwohl \n Advisors: \n Where: Bloomfield 424 and
ZOOM From:\nAriel University\nAbstract:\n\nThe classic herding model exami
nes the asymptotic behavior of agents who observe their predecessors' acti
ons as well as a private signal from an exogenous information structure. I
n this paper we introduce a self-interested sender into the model and stud
y her problem of designing this information structure. If agents cannot ob
serve each other the model reduces to Bayesian persuasion. However\, when
agents observe predecessors' actions\, they may learn from each other\, po
tentially harming the sender. We identify necessary and sufficient conditi
ons under which the sender can nevertheless obtain the same utility as whe
n the agents are unable to observe each other.\n\nJoint work with Itai Ari
eli and Rann Smorodinsky\n\nZoom Link\n\nhttps://technion.zoom.us/j/926243
64096
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:150@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20211229T113000
DTEND;TZID=Asia/Jerusalem:20211229T123000
DTSTAMP:20211226T095922Z
URL:https://dds.technion.ac.il/ar/iemevents/participatory-budgeting-via-a-
vcg-like-mechanism/
SUMMARY:Participatory Budgeting via a VCG-Like Mechanism [ \n Game Theo
ry Seminar\n Seminars\n \n ]
DESCRIPTION:By: Jonathan Wagner \n Advisors: \n Where: ZOOM From:\nTechnio
n\nAbstract:\n\nWe study a special divisible Participatory Budgeting model
that involves taxation decisions. Agents collectively choose the level of
tax collected from them to finance the city's public budget\, and in turn
the allocation of the resulting budget among different alternatives. We i
mplement a mechanism that is basically a VCG mechanism with some necessary
adjustments taking place (Most importantly\, our model violates the cruci
al assumption of quasi-linear utilities)\, that features stronger strategy
-proofness results as compared to existing divisible Participatory Budgeti
ng models. Moreover\, we specify the conditions that guarantee: (a) tha
t payments vanish with the population's size\, and (b) that non-negative p
ayments for all agents are feasible.\n\n \;\n\nJoint work with Reshef
Meir\n\n \;\n\nZoom Link\n\nhttps://technion.zoom.us/j/92624364096
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:153@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220105T113000
DTEND;TZID=Asia/Jerusalem:20220105T123000
DTSTAMP:20220104T102004Z
URL:https://dds.technion.ac.il/ar/iemevents/submodularity-in-interdependen
t-values/
SUMMARY:Submodularity in Interdependent Values [ \n Game Theory Seminar
\n Seminars\n \n ]
DESCRIPTION:By: DR. Alon Eden \n Advisors: \n Where: Bloomfield 424 and
ZOOM From:\nHarvard University\nAbstract:\n\nThe celebrated Interdependent
values (IDV) model captures settings where buyers have partial informatio
n regarding their value for the item being sold. In the IDV model\, eac
h buyer $i$ has a private piece of information\, a signal $s_i$\, and thei
r value is determined by all signals. This is represented by a valuation f
unction $v_i(s_1\, ...\, s_n)$. While this model is more realistic than th
e private values model (that assumes buyers know their value)\, it is much
less understood. Most works in the IDV model assume that there's only one
type of item being sold\, and that the valuation function $v_i$ is public
information.\n\n \;\n\nIn this talk\, I will present the first positi
ve results relaxing these assumptions. In order to obtain these results\,
we identify an interesting condition on the structure in which informatio
n affects valuations---submodularity over signals. Using this condition\,
we obtain a mechanism that achieves a 4-approximation to general combinato
rial auction settings\, and $O(log^2 n)$-approximation to a single item au
ction\, when the valuation $v_i$ is private information.\n\n \;\n\nBas
ed on joint works with Michal Feldman\, Amos Fiat\, Kira Goldner\, Anna Ka
rlin\, and Shuran Zheng.\n\nZoom Link\n\nhttps://technion.zoom.us/j/926243
64096
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:162@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220112T113000
DTEND;TZID=Asia/Jerusalem:20220112T123000
DTSTAMP:20220109T060152Z
URL:https://dds.technion.ac.il/ar/iemevents/populations-feasible-posterior
-beliefs/
SUMMARY:Population's Feasible Posterior Beliefs [ \n Game Theory Semina
r\n Seminars\n \n ]
DESCRIPTION:By: Prof. Yakov Babichenko \n Advisors: \n Where: ZOOM From:\n
Technion\nAbstract:\n\n \;\n\nWe consider a population of Bayesian age
nts who share a common prior over some finite state space and where each a
gent receives a private signal and forms a posterior belief. We ask which
distributions over aggregate joint posterior beliefs are feasible. We prov
ide a necessary and sufficient condition for such distribution to be feasi
ble. We study the implication of our result to a polarization of posterior
s in a population of agents. In addition\, we consider a private Bayesian
persuasion problem and\, using our characterization\, provide an optimal p
olicy for the sender.\n\n(Joint with Itai Arieli)\n\nZoom Link\nhttps://te
chnion.zoom.us/j/92624364096
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:172@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220302T113000
DTEND;TZID=Asia/Jerusalem:20220302T123000
DTSTAMP:20220224T062739Z
URL:https://dds.technion.ac.il/ar/iemevents/learning-to-cooperate-and-comp
ete-in-diplomacy/
SUMMARY:Learning to Cooperate and Compete in Diplomacy [ \n Game Theory
Seminar\n Seminars\n \n ]
DESCRIPTION:By: Noam Brown \n Advisors: \n Where: Bloomfield 424 From:\n
Facebook\nAbstract:\n\nAI has made incredible progress in purely adversari
al games such as chess\, go\, and poker. However\, the real world involves
a complex mixture of cooperation and competition\, sometimes with irratio
nal or suboptimal participants\, and in these settings past AI techniques
fall apart. For this reason\, Diplomacy\, a popular game focused on negoti
ation and alliance-building\, has served as a long-standing grand challeng
e for AI that requires radically different techniques compared to prior ga
mes and has major implications if AI algorithms eventually succeed. In thi
s talk I will describe Diplomacy and cover recent research results that ha
ve allowed our group to achieve expert-level human performance in the no-c
ommunication version of this game via a combination of imitation learning
on human data and self-play reinforcement learning. In particular\, I will
cover a new regret minimization technique that converges to a coarse corr
elated equilibrium regularized toward a human prior. I will conclude with
a discussion of the challenges that await as research shifts toward versio
ns of the game in which players can privately communicate and therefore co
rrelate their actions.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:186@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220330T113000
DTEND;TZID=Asia/Jerusalem:20220330T123000
DTSTAMP:20220325T053249Z
URL:https://dds.technion.ac.il/ar/iemevents/necessary-and-sufficient-condi
tions-for-determinacy-of-asymptotically-stationary-equilibria-in-olg-mode/
SUMMARY:Necessary and sufficient conditions for determinacy of asymptotical
ly stationary equilibria in olg mode [ \n Game Theory Seminar\n
Seminars\n \n ]
DESCRIPTION:By: Anna Rubinchik\n Advisors: \n Where: Bloomfield 424 and
Zoom Link From:\nWestern Galilee College\nAbstract:\n\nWe propose a crit
erion for verifying whether an equilibrium of an overlapping generation mo
del is amenable to local policy analysis\, i.e.\, is determinate. The crit
erion is applicable for a generic set of parameters of the model\, and in
case of indeterminacy\, it indicates the nature of the problem: multiplici
ty of equilibria or their absence for near-by parameters.\n\nThe criterion
can be applied to models with infinite past and future as well as those w
ith a truncated past.\n\nThe baseline equilibrium is not required to be a
steady state\, and economic parameters\, for example endowments\, can chan
ge over time.\n\nHowever\, asymptotically\, the equilibrium should be stat
ionary\, though the two limiting paths at either end of the time-line do n
ot have to be the same. If they are\, conditions for local uniqueness of e
quilibrium are far more stringent for an economy with a truncated past as
compared to its counterpart with an infinite past.\n\nWe illustrate our ma
in result using a text-book model with a single physical good and a two-pe
riod life-cycle. In this model our criterion is used to identify the three
possible cases: determinacy\, and hence\, local uniqueness\, and indeterm
inacy: multiplicity and non-existence of equilibria in the neighbourhood o
f the baseline. (The corresponding paper is here)\n\n \;\n\nZoom Lin
k \n\nhttps://technion.zoom.us/j/94175815347
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:191@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220406T113000
DTEND;TZID=Asia/Jerusalem:20220406T123000
DTSTAMP:20220401T103442Z
URL:https://dds.technion.ac.il/ar/iemevents/communicating-with-anecdotes-2
/
SUMMARY:Communicating with Anecdotes [ \n Game Theory Seminar\n Sem
inars\n \n ]
DESCRIPTION:By: Divyarthi Mohan\n Advisors: \n Where: Bloomfield 424 and Z
OOM From:\nTel Aviv University\nAbstract:\n\nWe study a communication game
between a sender and receiver where the sender has access to a set of inf
ormative signals about a state of the world. The sender chooses one of her
signals and communicates it to the receiver. We call this an ``anecdote''
. The receiver takes an action\, yielding a utility for both players. Send
er and receiver both care about the state of the world but are also influe
nced by a personal preference so that their ideal actions differ. We chara
cterize perfect Bayesian equilibria when the sender cannot commit to a par
ticular communication scheme. In this setting the sender faces ``persuasio
n temptation'': she is tempted to select a more biased anecdote to influen
ce the receiver's action. Anecdotes are still informative to the receiver
but persuasion comes at the cost of precision. This gives rise to ``inform
ational homophily'' where the receiver prefers to listen to like-minded se
nders because they provide higher-precision signals. We show that for fat-
tailed anecdote distributions the receiver might even prefer to talk to po
orly informed senders with aligned preferences rather than a knowledgeable
expert whose preferences may differ from her own because the expert's kno
wledge also gives her likely access to highly biased anecdotes. We also sh
ow that under commitment differences in personal preferences no longer aff
ect communication and the sender will generally report the most representa
tive anecdote closest to the posterior mean for common distributions.\n\nJ
oint work with Nika Haghtalab\, Nicole Immorlica\, Brendan Lucier\, and Ma
rkus Mobius.\n\nBe rational and attend.\n\n \;\n\nZoom Link\n\n \;
\n\nhttps://technion.zoom.us/j/94175815347
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:195@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220413T113000
DTEND;TZID=Asia/Jerusalem:20220413T123000
DTSTAMP:20220407T190719Z
URL:https://dds.technion.ac.il/ar/iemevents/learning-in-the-marriage-marke
t-the-economics-of-dating/
SUMMARY:Learning in the Marriage Market: The Economics of Dating [ \n G
ame Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Daniel Fershtman\n Advisors: \n Where: Bloomfield 424 and
ZOOM From:\nTel Aviv University\nAbstract:\n\nWe develop a dynamic model
of two-sided matching with search and learning frictions. Agents engage in
a search for a potential partner and\, upon meeting\, may gradually acqui
re information about their compatibility as a couple\, a process we refer
to as dating. Dating is mutually exclusive and\, as such\, introduces a tr
adeoff between becoming better informed about one's compatibility with a p
otential partner and meeting other\, more promising\, potential partners.
We derive a closed-form solution for the unique steady-state equilibrium w
hen agents are ex-ante homogeneous\, and characterize it when they are ver
tically heterogeneous. In the steady state\, agents date for longer than i
s socially optimal\, an inefficiency that is alleviated by a small degree
of asymmetry in dating costs between partners. Furthermore\, block segrega
tion fails\, yet matching is assortative -- in a probabilistic sense we re
fer to as single-crossing in marriage probabilities. Motivated by recent a
dvances in matching technologies in decentralized markets\, we study the e
ffects of improvements in search and learning technologies and show that t
hey differ qualitatively.\n\njoint with Yair Antler and Daniel Bird\n\nBe
rational and attend\,\n\nZoom Link\n\nhttps://technion.zoom.us/j/941758153
47
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:196@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220427T000000
DTEND;TZID=Asia/Jerusalem:20220427T000000
DTSTAMP:20220424T062033Z
URL:https://dds.technion.ac.il/ar/iemevents/dynamic-screening/
SUMMARY:Dynamic Screening [ \n Game Theory Seminar\n Seminars\n \
n ]
DESCRIPTION:By: David Lagziel \n Advisors: \n Where: Bloomfield 424 and
ZOOM From:\nBen-Gurion University\nAbstract:\n\nWe study dynamic screening
problems where elements are subjected to noisy evaluations and\, in every
stage\, some of the elements are rejected while the remaining ones are in
dependently re-evaluated in subsequent stages.\n\nWe prove that\, ceteris
paribus\, the quality of a dynamic screening process is not monotonic in t
he number of stages.\n\nSpecifically\, we examine the accepted elements' v
alues and show that adding a single stage to a screening process may produ
ce inferior results\, in terms of stochastic dominance\, whereas increasin
g the number of stages substantially leads to a first-best outcome.\n\nJoi
nt work with Ehud Lehrer\n\n \;\n\nZoom Link\n\nhttps://technion.zoom.
us/j/94175815347
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:207@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220511T113000
DTEND;TZID=Asia/Jerusalem:20220511T123000
DTSTAMP:20220508T053524Z
URL:https://dds.technion.ac.il/ar/iemevents/mentors-and-recombinators-mult
i-dimensional-social-learning/
SUMMARY:Mentors and Recombinators: Multi-Dimensional Social Learning [ \n
Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Yuval Heller \n Advisors: \n Where: Bloomfield 424 and Z
OOM From:\nBar-Ilan University\nAbstract:\n\n \;\n\nWe study games in
which the set of strategies is multi-dimensional\, and new agents might le
arn various strategic dimensions from different mentors. This induces a ne
w family of recombinator dynamics that generalises the standard replicator
dynamics\, in which each new agent learns everything from a single mentor
. We characterise the stable equilibria under these dynamics\, and we show
that they predict novel behaviour in various applications\, such as coope
ration supported by dominated strategies in the Prisoner's Dilemma\, and s
tability of robust traits at the expense of specialised traits.\n\n \;
\n\nJoint with Omer Edhan\, Ziv Hellman\, Srinivas Arigapudi.\n\n\nZoom Li
nk\n\nhttps://technion.zoom.us/j/94175815347
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:211@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220518T113000
DTEND;TZID=Asia/Jerusalem:20220518T123000
DTSTAMP:20220515T045358Z
URL:https://dds.technion.ac.il/ar/iemevents/calibrated-forecasting-and-per
suasion/
SUMMARY:Calibrated Forecasting and Persuasion [ \n Game Theory Seminar\
n \n ]
DESCRIPTION:By: Atulya Jain \n Advisors: \n Where: Bloomfield 424 and ZOO
M From:\nHEC Paris\nAbstract:\n\n \;\n\nHow should an expert send fore
casts to maximize her payoff subject to passing the calibration test? We c
onsider a dynamic game where an expert provides probability forecasts to a
decision maker. The decision maker uses past outcomes to verify the claim
s of the expert using the calibration test. Under reasonable assumptions\,
we characterize the dynamic forecasting game in terms of a static persuas
ion problem. In doing so\, we determine the optimal calibrated forecasting
strategy. We compare the attainable payoffs for an informed and an uninfo
rmed expert. Finally\, we compare the analysis when the decision maker is
a regret minimizer. We show that the expert can always guarantee the cali
bration benchmark and\, in some instances\, she can guarantee strictly mor
e.\n\n \;\n\nJoint with with Vianney Perchet\n\nZoom Link\n\nhttps://t
echnion.zoom.us/j/94175815347
CATEGORIES:Game Theory Seminar
END:VEVENT
BEGIN:VEVENT
UID:214@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220525T113000
DTEND;TZID=Asia/Jerusalem:20220525T123000
DTSTAMP:20220522T081142Z
URL:https://dds.technion.ac.il/ar/iemevents/monopoly-product-quality-and-f
lexible-learning/
SUMMARY:Monopoly\, Product Quality\, and Flexible Learning [ \n Game Th
eory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Jeffrey Mensch \n Advisors: \n Where: Bloomfield 424 and
ZOOM From:\nThe Hebrew University\nAbstract:\n\n \;\n\nA seller offer
s a buyer a schedule of transfers and associated product qualities\, as in
Mussa and Rosen (1978). After observing this schedule\, the buyer chooses
a flexible costly signal about his type. We show it is without loss to fo
cus on a class of mechanisms that compensate the buyer for his learning co
sts. Using these mechanisms\, we prove quality always lies strictly below
the efficient level. This strict downward distortion holds even if the buy
er acquires no information or when the buyer’s posterior type is the hig
hest possible given his signal\, reversing the “no distortion at the top
” feature that holds when information is exogenous.\n\n \;\n\nJoint
with with Doron Ravid\n\n \;\n\nZoom Link\n\nhttps://technion.zoom.us/
j/94175815347
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:220@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220601T131500
DTEND;TZID=Asia/Jerusalem:20220601T141500
DTSTAMP:20220529T100848Z
URL:https://dds.technion.ac.il/ar/iemevents/a-combinatorial-characterizati
on-of-minimax-in-0-1-games/
SUMMARY:A Combinatorial Characterization of Minimax in 0/1 Games [ \n G
ame Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Shay Moran \n Advisors: \n Where: Bloomfield 424 and ZOO
M From:\nThe Technion\nAbstract:\n\nWe will discuss a generalization of th
e celebrated Minimax Theorem (von Neumann\, 1928) for binary zero-sum game
s. A simple game which fails to satisfy Minimax is Ephraim Kishon's “Jew
ish Poker” (see [1\,2] below). In this game\, each player picks a number
and the larger number wins. The payoff matrix in this game is *infinite t
riangular*. We show this is the only obstruction: if a game does not conta
in triangular submatrices of unbounded sizes then the Minimax Theorem hold
s. This generalizes von Neumann's Minimax Theorem by removing requirements
of finiteness or compactness.\n\nJoint work with Steve Hanneke and Roi Li
vni.\n\n[1] http://www.ephraimkishon.de/en/my_favorite_stories.htm (englis
h)\n[2] https://gesherfilmfund.org.il/documents/מקבץ%20יצירות%20
%20-%20אפרים%20קישון.pdf (hebrew\, third story)\n\n \;\n\nZo
om Link\n\nhttps://technion.zoom.us/j/94175815347
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:227@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220615T113000
DTEND;TZID=Asia/Jerusalem:20220615T123000
DTSTAMP:20220807T120338Z
URL:https://dds.technion.ac.il/ar/iemevents/the-effect-of-partial-vertical
-ownership-on-operational-decisions-and-information-exchange-in-a-supply-c
hain/
SUMMARY:The effect of Partial Vertical Ownership on Operational Decisions a
nd Information Exchange in a Supply Chain [ \n Game Theory Seminar\n
Seminars\n \n ]
DESCRIPTION:By: Tal Avinadav \n Advisors: \n Where: Bloomfield 424 and Z
OOM From:\nBar Ilan University\nAbstract:\n\nPartial vertical ownership de
scribes a situation in which a firm holds financial shares either in its s
upplier (referred to as partial backward integration) or in its customer (
partial forward integration). We study the effect of such financial interc
onnectedness on two operational decisions: capacity investment and informa
tion exchange. In our model\, a retailer\, who has superior information ab
out the future market demand\, possesses some level of passive financial h
oldings in the supplier. Although this passive financial investment does n
ot enable the retailer to directly influence the supplier’s operational
decisions\, it does affect the market equilibrium. Specifically\, financia
l interconnectedness between the firms can result in the retailer financin
g the entire capacity in the market. In addition\, we characterize the con
ditions that ensure information between the retailer and the supplier can
be exchanged by “cheap-talk” communication. When “cheap talk” is n
ot possible\, we study the separating equilibrium that is achieved by the
retailer’s commitment to order in advance. Interestingly\, this advance
order can be either decreasing or increasing with respect to the level of
the financial holdings.\n\n \;\n\nZoom Link\n\nhttps://technion.zoom.u
s/j/94175815347\n\n \;\n\n \;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:231@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220622T113000
DTEND;TZID=Asia/Jerusalem:20220622T123000
DTSTAMP:20220807T124445Z
URL:https://dds.technion.ac.il/ar/iemevents/envy-driven-equilibrium-in-sin
gle-peaked-single-crossing-cheap-talk-games/
SUMMARY:Envy-driven equilibrium in single-peaked\, single-crossing cheap ta
lk games [ \n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Françoise Forges\n Advisors: \n Where: ZOOM From:\nParis
Dauphine University\nAbstract:\n\nIn cheap talk games\, equilibrium refine
ments usually start from a supposedly played equilibrium\, and consider th
e incentives of some types to separate off path from the types they are po
oled with. We take the opposite direction. We propose algorithms that star
t from a fully separating strategy\, and merge types until an equilibrium
is reached. These algorithms were first conceived to establish the existen
ce of a perfect Bayesian equilibrium in sender-receiver games with sender
’s approval in which the utility functions are single-peaked and single-
crossing. When the sender’s approval is not an issue\, our algorithms ac
hieve a unique equilibrium\, which fulfills many standard refinement crite
ria.\n\n \;\n\nZoom Link\n\nhttps://technion.zoom.us/j/94175815347
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:259@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20220921T113000
DTEND;TZID=Asia/Jerusalem:20220921T123000
DTSTAMP:20220915T063325Z
URL:https://dds.technion.ac.il/ar/iemevents/algorithmic-contract-theory/
SUMMARY:Algorithmic Contract Theory [ \n Game Theory Seminar\n Semi
nars\n \n ]
DESCRIPTION:By: Elisheva Shamash \n Advisors: \n Where: Bloomfield 424 F
rom:\nTechnion\nAbstract:\n\nWe consider contracts with common agencies st
emming from the VCG-mechanism. These contracts have strategic common agent
s that take actions on behalf of multiple principals\, while aiming to max
imizes their personal profits. In these contracts the agent takes the role
of the Mechanism Designer (MD) in the VCG-mechanism\, but also acts strat
egically to maximize his private utility.\n\nMost contracts in practice an
d the literature are based on classic contracts\, where the principals ann
ounce prices for possible outcomes\, and pay their announced price for the
realized outcome. However\, classic contracts may lead to extremely ineff
icient outcomes.\n\nOur first paper Principal Agent VCG Contracts (Lavi an
d Shamash 2022) introduces a novel contract with complete information term
ed a VCG-contract. The structure of the VCG-contract induces M principals
to announce their true values for agent actions\, and N agents to choose a
ctions similar to those acquired in a VCG-auction\, where in the case of a
single agent maximize social welfare.\nWe characterize the set of instanc
es inducing pure sub-game perfect equilibria (SPE) and show that the worst
case efficiency result in VCG-contracts is not worse than that of classic
contracts\, when agents have over two actions.\n\nOur second paper Incomp
lete Information VCG Contracts for Common Agency (Alon et al 2021) general
izes a single agent case of Lavi and Shamash 2022 by analyzing the incompl
ete information setting - where each of the agent's actions leads to a dis
tribution over outcomes\, and principals only perceive the resulting stoch
astic outcome of the chosen action.\nWe characterize the unique class of i
ncomplete information VCG-contracts (IIVCG) that induce principals to anno
unce their true outcome values\, and the agent to choose the actions that
maximizes expected social welfare.\nIn addition\, we show that no IIVCG co
ntract exists\, that guarantees both individual rationality (IR) and limit
ed liability (LL).\nFurthermore\, we develop a computationally efficient a
lgorithm determining whether\, given a specific setting\, a contract belon
ging to IIVCG satisfying both LL and IR exists\, and if so generates such
a contract.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:271@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20221109T113000
DTEND;TZID=Asia/Jerusalem:20221109T123000
DTSTAMP:20221106T115556Z
URL:https://dds.technion.ac.il/ar/iemevents/equity-crowdfunding/
SUMMARY:Equity Crowdfunding [ \n Game Theory Seminar\n Seminars\n
\n ]
DESCRIPTION:By: Ella Segev \n Advisors: \n Where: Room 424\, Bloomfield bu
ilding From:\nHebrew University\nAbstract: Equity crowdfunding is a met
hod of financing an initiative whereby an entrepreneur sells shares in her
firm to a group of people (the crowd) on a dedicated platform. Since 2015
\, when equity crowdfunding became legal in the USA under regulation A+\,
it has become increasingly popular\, and its potential economic impact is
significant. Understanding the forces that shape the behavior of both inv
estors and entrepreneurs in equity crowdfunding platforms can help design
more efficient platforms and increase the welfare of all participants. We
therefore develop a common value sequential crowdfunding game-theoretic mo
del\, where the entrepreneur sells a percentage of her firm in order to ra
ise money for its establishment and then shares the future value of the fi
rm with the crowd. Investors on the platform who visit the campaign decide
whether or not to invest in it. Each investor’s decision depends on the
amount that has already been invested before him and on his own knowledge
about the firm and the market in which it operates (which we model as a s
ignal that he obtains regarding the true value of the firm). By offering a
different percentage in the firm\, the entrepreneur leads the crowd to a
different equilibrium. We characterize these equilibria and then analyze t
he entrepreneur’s decision. We show that an entrepreneur with a higher e
x-ante probability of success (a better firm) will offer a higher percenta
ge in the firm (or equivalently a lower price per share) in order to incre
ase the probability that the campaign succeeds. We further show that the e
ntrepreneur may very well prefer investors that have a less accurate signa
l regarding the true value of the firm.\n \;\n\nJoint work with Hana
Tzur
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:275@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20221116T113000
DTEND;TZID=Asia/Jerusalem:20221116T123000
DTSTAMP:20221113T114747Z
URL:https://dds.technion.ac.il/ar/iemevents/the-effect-of-privacy-on-marke
t-structure-and-prices/
SUMMARY:The Effect of Privacy on Market Structure and Prices. [ \n Game
Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Daniel Bird \n Advisors: \n Where: Bloomfield 424 From:\nT
el Aviv University\nAbstract:\nProtection of consumers’ privacy is often
motivated by the fear that\, without it\, consumers may be exploited via
personalized pricing. However\, personalized pricing seems to be rare in p
ractice. We explain how privacy nevertheless affects prices in search mark
ets through its effect on market structure. If privacy is not protected\,
then in addition to consumer search\, firms may engage in targeted adverti
sing. We show that privacy protection reduces consumer surplus if firms pr
ice discriminate between the search and advertising markets. Absent such d
iscrimination\, privacy protection may either increase or decrease consume
r surplus. We relate our results to the “privacy paradox.”
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:278@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20221123T113000
DTEND;TZID=Asia/Jerusalem:20221123T123000
DTSTAMP:20221123T070620Z
URL:https://dds.technion.ac.il/ar/iemevents/learning-by-underfitting-and-t
he-power-of-regret-minimization/
SUMMARY:Learning by underfitting and the power of regret minimization [ \n
Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Tomer Koren \n Advisors: \n Where: Bloomfield 526 From:\
nTel Aviv University\nAbstract: In this talk\, I will discuss how the g
ame-theoretic concept of regret minimization plays a pivotal role towards
generalization of modern optimization methods in machine learning. I wil
l illustrate this through some curious theoretical results and phenomena c
oncerning Stochastic Gradient Descent (SGD) that challenge the current con
ventional wisdom in machine learning. The talk will not assume any prior
knowledge in machine learning and/or optimization.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:286@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20221207T113000
DTEND;TZID=Asia/Jerusalem:20221207T123000
DTSTAMP:20221206T105110Z
URL:https://dds.technion.ac.il/ar/iemevents/fair-shares-feasibility-domina
tion-and-incentives/
SUMMARY:Fair Shares: Feasibility\, Domination and Incentives [ \n Game
Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Moshe Babaioff \n Advisors: \n Where: Bloomfield 527 From:
\nMicrosoft Research\nAbstract:\n\nWe consider the problem of fairly alloc
ating a set of indivisible goods to equally entitled agents\, with no mone
tary transfers. A share function maps a pair of agent valuation and number
of agents to a non-negative value\, with the interpretation that if an al
location fails to give the agent a bundle of value at least equal to her s
hare\, this serves as evidence that the allocation is not fair towards the
agent. We embark on a systematic study of feasible shares – a share is
feasible if we can always give every agent her share. We introduce the not
ion of a self-maximizing share to capture the incentive for truthful valua
tion reporting. We prove the inexistence of an “ultimate feasible shar
e”: one that dominates any other feasible share that is self-maximizing.
We then present several feasible shares that are self-maximizing and poly
nomial-time computable\, and approximately dominate all other shares.\n\n&
nbsp\;\nJoint with Uri Feige
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:299@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20221214T113000
DTEND;TZID=Asia/Jerusalem:20221214T123000
DTSTAMP:20221213T074353Z
URL:https://dds.technion.ac.il/ar/iemevents/the-geometry-of-consumer-prefe
rence-aggregation/
SUMMARY:The geometry of consumer preference aggregation [ \n Game Theor
y Seminar\n Seminars\n \n ]
DESCRIPTION:By: Fedor Sandomirskiy \n Advisors: \n Where: Bloomfield 424
From:\nCalifornia Institute of Technology\nAbstract:\n\nThis paper revisit
s a classical question in economics: how individual preferences and income
s shape aggregate behavior. We develop a method that reduces the hard prob
lem of aggregation to simply computing a weighted average. The method appl
ies to populations with homothetic preferences. The key idea is to handle
aggregation in the space of logarithmic expenditure functions.\n\nWe demon
strate the power of this method by (i) characterizing classes of preferenc
es invariant with respect to aggregation\, i.e.\, such that any population
of heterogeneous consumers with preferences from the class behaves as if
it were a single aggregate consumer from the same class\; (ii) characteriz
ing classes of aggregate preferences generated by popular preference domai
ns such as linear or Leontief\; (iii) describing indecomposable preference
s\, i.e.\, those that do not correspond to aggregate behavior of any non-t
rivial population\; (iv) representing any preference as an aggregation of
indecomposable ones.\n\nWe discuss connections and applications of our fin
dings to stochastic discrete choice\, information design\, welfare analysi
s and gains from trade estimation\, pseudo-market mechanisms\, and prefere
nce identification.\n\n \;\nJoint with Philip Ushchev
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:302@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20221221T113000
DTEND;TZID=Asia/Jerusalem:20221221T123000
DTSTAMP:20221219T070544Z
URL:https://dds.technion.ac.il/ar/iemevents/contract-design-in-combinatori
al-settings/
SUMMARY:Contract design in combinatorial settings [ \n Game Theory Semi
nar\n Seminars\n \n ]
DESCRIPTION:By: Tomer Ezra \n Advisors: \n Where: Bloomfield 424 From:\nSa
pienza University of Rome\nAbstract:\n\nWe study two combinatorial setting
s of the contract design problem\, in which a principal wants to delegate
the execution of a costly task. In the first setting\, the principal deleg
ates the task to an agent that can take any subset of a given set of unobs
ervable actions\, each of which has an associated cost. The principal rece
ives a reward which is a combinatorial function of the actions taken by th
e agent.\n\nIn the second setting\, we study the single-principal multi-ag
ent contract problem\, in which the principal motivates a team of agents t
o exert effort toward a given task.\n\nWe design (approximately) optimal a
lgorithms for both settings along with impossibility results for various c
lasses of combinatorial functions.\n\nIn particular\, for the single agent
setting\, we show that if the reward function is gross substitutes\, then
an optimal contract can be computed with polynomially many value queries\
, whereas if it is submodular\, the optimal contract is NP-hard. For the m
ulti-agent setting\, we show how using demand and value queries\, it is po
ssible to obtain a constant approximation\, where for subadditive reward f
unctions it is impossible to achieve an approximation of o(\\sqrt(n)).\n\n
Our analysis uncovers key properties of gross substitutes and XOS function
s\, and reveals many interesting connections between combinatorial contrac
ts and combinatorial auctions.\nJoint work with Paul Duetting\, Michal Fel
dman\, and Thomas Kesselheim
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:305@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20221228T113000
DTEND;TZID=Asia/Jerusalem:20221228T123000
DTSTAMP:20221228T064613Z
URL:https://dds.technion.ac.il/ar/iemevents/evolutionary-dynamics-in-bilin
gual-games/
SUMMARY:: Evolutionary Dynamics in Bilingual Games [ \n Game Theory Se
minar\n Seminars\n \n ]
DESCRIPTION:By: Srinivas Arigapudi \n Advisors: \n Where: Bloomfield 424
From:\nTechnion\nAbstract:\n\nIn two-strategy coordination games with dist
inct payoff- and risk-dominant equilibria\, existing results show that the
inefficient risk-dominant equilibrium is uniquely selected in the infinit
e horizon under many evolutionary dynamics. In the above class of coordina
tion games\, we study the effect of introducing a bilingual strategy that
is compatible with both of the existing strategies. An agent playing the b
ilingual strategy incurs an additional adoption cost but never miscordinat
es with any other agent. We show that if the adoption cost of the bilingua
l strategy is low\, then the efficient payoff-dominant equilibrium can be
uniquely selected not only in the infinite horizon but also in the finite
horizon under many evolutionary dynamics.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:309@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230104T113000
DTEND;TZID=Asia/Jerusalem:20230104T123000
DTSTAMP:20230102T063037Z
URL:https://dds.technion.ac.il/ar/iemevents/auctions-between-regret-minimi
zing-agents-2/
SUMMARY:Auctions Between Regret-Minimizing Agents [ \n Game Theory Semi
nar\n Seminars\n \n ]
DESCRIPTION:By: Noam Nisan \n Advisors: \n Where: Bloomfield 424 From:\nHe
brew University \nAbstract:\n\nWe analyze a scenario in which software age
nts implemented as regret-minimizing algorithms engage in a repeated aucti
on on behalf of their users. We study first price and second price auction
s\, as well as their generalized versions (e.g.\, as those used for ad auc
tions). Using both theoretical analysis and simulations\, we show that\, s
urprisingly\, in second price auctions the players have incentives to mis-
report their true valuations to their own learning agents\, while in the f
irst price auction it is a dominant strategy for all players to truthfully
report their valuations to their agents.\n\n \;\n\nSee paper on arXiv
: https://arxiv.org/abs/2110.11855\, as well as a related companion paper:
https://arxiv.org/abs/2112.07640.\n\n \;\nJoint work with Yoav Kolum
bus.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:330@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230322T113000
DTEND;TZID=Asia/Jerusalem:20230322T123000
DTSTAMP:20230314T083918Z
URL:https://dds.technion.ac.il/ar/iemevents/bargaining-over-treatment-choi
ce-under-disagreement/
SUMMARY:Bargaining over Treatment Choice under Disagreement [ \n Game T
heory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Nabil I. Al-Najjar\n Advisors: \n Where: Bloomfield 424 Fr
om:\nNorthwestern University\nAbstract:\n \nA group of experts with diffe
rent prior beliefs must choose a treatment. A dataset is made public and l
eads to revisions of beliefs. We propose a model where the experts’ disa
greements are resolved through bargaining\, using the Nash bargaining solu
tion. Experts bargain after disclosure of the dataset. Bargaining may lead
to an inefficient use of information in a strong sense: experts receive a
lower payoff in every state\, and for any prior belief (i.e.\, inadmissib
ility). Bargaining exhibits under-reaction to information as compared to t
he normative solution in which experts bargain ex ante on the procedure us
ed to exploit the data.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:327@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230329T113000
DTEND;TZID=Asia/Jerusalem:20230329T123000
DTSTAMP:20230308T060002Z
URL:https://dds.technion.ac.il/ar/iemevents/auctions-with-frictions/
SUMMARY:Auctions with Frictions [ \n Game Theory Seminar\n Seminars
\n \n ]
DESCRIPTION:By: Asher Wolinsky\n Advisors: \n Where: Bloomfield 427 From:\
nNorth Western\nAbstract:\n\nThe "auction with frictions" model features c
ostly recruiting of bidders\, costly bidders' participation and limited se
ller's commitment ability. The latter failure is probably the main qualita
tive departure of our analysis from prior work that features costly recrui
tment or entry costs. It affects both the recruitment and the bidding phas
es: in the former phase\, the seller may be unable to commit to the extent
of the recruitment effort\; in the latter phase\, the seller might not be
able to credibly disclose the extent of the actual participation or commi
t to a positive reserve price.\n\nOne set of insights concerns inherent in
efficiencies: excessive recruitment effort and inefficient extent of entry
. The latter sometimes takes the extreme form of complete shutdown of trad
e\, even when the potential gains are non-negligible. Another set of insig
hts concerns the effect of bidders' ability to observe the extent of parti
cipation (or seller's ability to credibly disclose it). It identifies cond
itions under which observability can be beneficial or detrimental for trad
e. Yet another set of insights concerns revenue ranking of the first and s
econd price auction formats. In the presence of these frictions\, they are
not equivalent\, and their ranking depends on recruitment and entry-cost
conditions.\n\nA later part of the paper examines how seller's private inf
ormation may affect this interaction. One of the insights here is that the
market may unravel almost completely. Almost all seller types may stay ou
t of the market (do not recruit bidders)\, not realizing potential gains f
rom trade.\n\nThese insights do not arise in the existing literature on au
ctions with recruitment and entry costs\, where the seller can commit to p
articular participation levels or use entry fees or subsidies to ensure op
timal entry.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:339@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230419T113000
DTEND;TZID=Asia/Jerusalem:20230419T123000
DTSTAMP:20230416T121844Z
URL:https://dds.technion.ac.il/ar/iemevents/optimal-insurance-dual-utility
-random-losses-and-adverse-selection/
SUMMARY:Optimal Insurance: Dual Utility\, Random Losses and Adverse Selecti
on [ \n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Alex Gershkov \n Advisors: \n Where: Bloomfield 424 From:\
nThe Hebrew University \nAbstract: \n\nWe study a generalization of the
classical monopoly insurance problem under adverse selection (see Stiglit
z [1977]) where we allow for a random distribution of losses\, possibly co
rrelated with the agent’s risk parameter that is private information. Ou
r main purpose is to provide a convenient analytic model that explains bot
h the pattern of observed customer behavior and the pattern of insurance c
ontracts most often observed in practice: these consist of menus of severa
l deductible-premium pairs\, or menus of insurance with coverage limits-pr
emium pairs. The main departure from the classical insurance literature is
obtained here by endowing the agents with risk-averse preferences that ca
n be represented by a dual utility functional (Yaari [1987]).\n\n \;\n
\njoint work with B. Moldovanu\, P. Strack and M. Zhang
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:338@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230419T133000
DTEND;TZID=Asia/Jerusalem:20230419T143000
DTSTAMP:20230416T121911Z
URL:https://dds.technion.ac.il/ar/iemevents/reputation-based-persuasion-pl
atforms/
SUMMARY:Reputation-based Persuasion Platforms [ \n Game Theory Seminar\
n Seminars\n \n ]
DESCRIPTION:By: Omer Madmon \n Advisors: \n Where: Bloomfield 526 From:\nT
echnion\nAbstract:\n\nIn this paper we introduce a two-stage Bayesian pers
uasion model in which a third-party platform controls the information avai
lable to the sender about users’ preferences. We aim to characterize the
optimal information disclosure policy of the platform\, which maximizes a
verage user utility\, under the assumption that the sender also follows it
s own optimal policy. We show that this problem can be reduced to a model
of market segmentation\, in which probabilities are mapped into valuations
. We then introduce a repeated variation of the persuasion platform proble
m in which myopic users arrive sequentially. In this setting\, the platfor
m controls the sender’s information about users and maintains a reputati
on for the sender\, punishing it if it fails to act truthfully on a certai
n subset of signals. We provide a characterization of the optimal platform
policy in the reputation-based setting\, which is then used to simplify t
he optimization problem of the platform.\n\n \;\n\nJoint work with M.
Tennenholtz\, and I. Arieli
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:346@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230503T113000
DTEND;TZID=Asia/Jerusalem:20230503T123000
DTSTAMP:20230502T075616Z
URL:https://dds.technion.ac.il/ar/iemevents/informationally-robust-cheap-t
alk/
SUMMARY:Informationally Robust Cheap-Talk [ \n Game Theory Seminar\n
Seminars\n \n ]
DESCRIPTION:By: Ronen Gradwohl \n Advisors: \n Where: Bloomfield 424 From:
\nAriel University\nAbstract:\n\n \;\n\nWe study the robustness of che
ap-talk equilibria to infinitesimal private information of the receiver in
a model with a binary state-space and state-independent sender-preference
s. We show that the sender-optimal equilibrium is robust if and only if th
is equilibrium either reveals no information to the receiver or fully reve
als one of the states with positive probability. We then characterize the
actions that can be played with positive probability in any robust equilib
rium. Finally\, we fully characterize the optimal sender-utility under bin
ary receiver's private information\, and provide bounds for the optimal se
nder-utility under general private information.\n\n \;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:349@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230511T113000
DTEND;TZID=Asia/Jerusalem:20230511T123000
DTSTAMP:20230509T102959Z
URL:https://dds.technion.ac.il/ar/iemevents/allocation-mechanisms-with-mix
ture-averse-preferences/
SUMMARY:Allocation Mechanisms with Mixture-Averse Preferences [ \n Game
Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Uzi Segal\n Advisors: \n Where: Bloomfield 424 From:\nBost
on College\nAbstract:\n\nWe study a variant of the house allocation proble
m (one-sided matching). Consider a continuum economy where N types of good
s need to be allocated\, one for each agent. All agents agree on the ranki
ng of the basic goods and have quasi-convex preferences over lotteries. An
optimal solution is a feasible allocation that is Pareto efficient and sa
tisfies No Envy. We show that an optimal solution must give each agent a b
inary lottery\, but identical agents may receive different lotteries. We p
rove the existence of an optimal solution when all individuals have the sa
me preferences over lotteries\, and provide sufficient conditions for exis
tence if individuals' preferences are ranked in terms of risk aversion. If
individuals satisfy the reduction of compound lotteries axiom\, then the
social planner can only deteriorate welfare by first randomizing over thes
e binary lotteries\, hence giving all agents the same lottery is never opt
imal. Full ex-ante equality can be achieved if agents satisfy the compound
independence axiom.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:350@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230517T113000
DTEND;TZID=Asia/Jerusalem:20230517T123000
DTSTAMP:20230517T055901Z
URL:https://dds.technion.ac.il/ar/iemevents/certification-markets-with-sig
naling-considerations/
SUMMARY:Certification Markets with Signaling Considerations [ \n Game T
heory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Ran Weksler \n Advisors: \n Where: Bloomfield 424 From:\n
Haifa University\nAbstract: \n\nIn markets for tests\, as the cost for co
nsumers plays a signaling role\, prices and grades interact to determine t
he market’s inference of the consumers’ values\, which in turn affects
the consumers’ willingness to pay for the test. In this paper\, we stud
y how the signaling role of prices affects the equilibrium outcome in mark
ets for tests - competition\, and monopoly. We show that although the exte
nt of the equilibrium signaling is not correlated with a test’s informat
iveness level\, a monopolist chooses the least informative test. In contra
st\, competing certifiers choose the most informative one. Nonetheless\, s
ince pricing strategy also releases information to the market\, it could b
e that the amount of information released in equilibrium would be higher u
nder the monopolist regime.\n\n \;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:352@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230524T113000
DTEND;TZID=Asia/Jerusalem:20230524T123000
DTSTAMP:20230524T051713Z
URL:https://dds.technion.ac.il/ar/iemevents/are-simple-mechanisms-optimal-
when-agents-are-unsophisticated/
SUMMARY:Are Simple Mechanisms Optimal when Agents are Unsophisticated? [ \n
Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Jiangtao Li \n Advisors: \n Where: Bloomfield 424 From:\nS
ingapore Management University\nAbstract:\n\nWe study the design of mechan
isms involving agents that have limited strategic sophistication. We defin
e a mechanism to be simple if---given the assumed level of strategic sophi
stication---agents can determine their optimal strategy. We examine whethe
r it is simple for the mechanism designer who faces strategically unsophis
ticated agents to offer a simple mechanism. We show that when the designer
uses a mechanism that is not simple\, while she loses the ability to pred
ict play\, she may nevertheless be better off no matter how agents resolve
their strategic confusion.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:355@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230531T000000
DTEND;TZID=Asia/Jerusalem:20230531T000000
DTSTAMP:20230530T110729Z
URL:https://dds.technion.ac.il/ar/iemevents/does-the-model-of-voting-in-a-
stochastic-environment-help-to-better-understand-real-voting/
SUMMARY:Does the model of voting in a stochastic environment help to better
understand real voting? [ \n Game Theory Seminar\n Seminars\n \
n ]
DESCRIPTION:By: Pavel Chebotarev \n Advisors: \n Where: Bloomfield 424 Fro
m:\nTechnion \nAbstract: \n\nIn the ViSE (Voting in Stochastic Environmen
t) model\, proposals for voting are realizations of a multivariate random
variable. Each proposal is a vector of voters’ gains/losses. The study o
f this model helps to understand how a society whose size\, decision-makin
g procedures\, and a set of individual voting algorithms are given\, will
succeed in a stochastic environment with given properties. This gives soci
ety and voters a chance to optimize their parameters. The talk presents se
veral results of the study of the model\, some of which are somewhat surpr
ising.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:367@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20230614T133000
DTEND;TZID=Asia/Jerusalem:20230614T140000
DTSTAMP:20230612T090535Z
URL:https://dds.technion.ac.il/ar/iemevents/blotto-game-with-testing-the-l
ocks-bombs-and-testing-model/
SUMMARY:Blotto Game with Testing (The Locks\, Bombs and Testing Model) [ \n
Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Issac Sonin \n Advisors: \n Where: Bloomfield 424 From:\nU
niversity of North Carolina \nAbstract:\n\nWe present a Defense/Attack res
ource allocation model\, where the Defender has some number of “locks”
to protect n vulnerable boxes (sites)\, and the Attacker is trying to des
troy these boxes\, having some number of “bombs” which can be placed i
nto the boxes. Similar models were studied in game theory - e.g. (Colonel)
Blotto games\, but our model has a feature that is absent in the previous
literature. Here\, the Attacker tests the vulnerability of all sites befo
re allocating her resources. However\, these tests are not perfect\, i.e.\
, a test can give a positive result for a box without a lock and a negativ
e result for a box with a lock. We give a full solution for some important
special cases of a general Locks-Bombs-Testing (G-LBT) model\, where site
s have distinct values. These results complement previously known results
for a symmetrical LBT model\, where all parameters are the same for each b
ox. We also outline an approach to a solution for the G-LBT model.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:413@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240117T113000
DTEND;TZID=Asia/Jerusalem:20240117T123000
DTSTAMP:20240114T075901Z
URL:https://dds.technion.ac.il/ar/iemevents/identifying-the-deviator/
SUMMARY:Identifying the Deviator [ \n Game Theory Seminar\n Seminar
s\n \n ]
DESCRIPTION:By: Eran Shmaya \n Advisors: \n Where: Bloomfield 424 From:\
nStony Brook University\nAbstract: A group of players are supposed to foll
ow a prescribed profile of strategies. If they follow this profile\, they
will reach a given target. We show that if the target is not reached becau
se some player deviates\, then an outside observer can identify the deviat
or. We also construct identification methods in two nontrivial cases.\n\n&
nbsp\;\n\nJoint work with Noga Alon\, Benjamin Gunby\, Xiaoyu He\, and Eil
on Solan\n\n \;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:417@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240124T113000
DTEND;TZID=Asia/Jerusalem:20240124T123000
DTSTAMP:20240123T072506Z
URL:https://dds.technion.ac.il/ar/iemevents/algorithmic-cheap-talk/
SUMMARY:Algorithmic Cheap Talk [ \n Game Theory Seminar\n Seminars\
n \n ]
DESCRIPTION:By: Konstantin Zabarnyi \n Advisors: \n Where: Bloomfield 424
From:\nTechnion\nAbstract: \n\n \;\n\nThe literature on strategic co
mmunication originated with the influential cheap talk model\, which prece
des the Bayesian persuasion model by three decades. This model describes a
n interaction between two agents: sender and receiver. The sender knows so
me state of the world that the receiver does not know and tries to influe
nce the receiver's action by communicating a cheap talk message to the rec
eiver.\n\nThis paper initiates the algorithmic study of cheap talk in a fi
nite environment (i.e.\, a finite number of states and the receiver's poss
ible actions). We first prove that approximating the sender-optimal or the
welfare-maximizing cheap talk equilibrium up to a certain additive consta
nt or multiplicative factor is NP-hard. Fortunately\, we identify three na
turally-restricted cases with efficient algorithms for finding a sender-op
timal equilibrium. These include a state-independent sender's utility stru
cture\, a constant number of states\, or a receiver having only two action
s.\n\n \;\nJoint work with Yakov Babichenko\, Inbal Talgam-Cohen\, and
Haifeng Xu
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:421@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240131T113000
DTEND;TZID=Asia/Jerusalem:20240131T123000
DTSTAMP:20240129T090556Z
URL:https://dds.technion.ac.il/ar/iemevents/the-indoctrination-game/
SUMMARY:The Indoctrination Game [ \n Game Theory Seminar\n Seminars
\n \n ]
DESCRIPTION:By: David Lagziel \n Advisors: \n Where: Bloomfield 424 From:
\nBen-Gurion University\nAbstract: \n\n \;\n\nThe indoctrination game
is a full-information contest over public opinion. The players exert cost
ly effort to publicly express their private opinions\, striving to dominat
e the discourse and thereby steer the prevailing opinion to align with th
eir own. Our research provides a theoretical basis for the phenomena of th
e silent majority and vocal minority. We demonstrate that\, in equilibrium
\, moderate opinions are suppressed\, thereby affording extremists unbridl
ed control over the dialogue. Furthermore\, we show that heightened exposu
re to diverse perspectives escalates the perceptible polarization within a
given population. Drawing on these insights\, we formulate a new social
-learning framework\, referred to as 'an indoctrination process'. Consis
tent with empirical evidence\, our findings forecast a monotonic escalatio
n in polarization as societal interconnectedness intensifies.\n\n \;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:423@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240207T113000
DTEND;TZID=Asia/Jerusalem:20240207T123000
DTSTAMP:20240206T135526Z
URL:https://dds.technion.ac.il/ar/iemevents/task-allocation-in-networks/
SUMMARY:Task Allocation in Networks [ \n Game Theory Seminar\n Semi
nars\n \n ]
DESCRIPTION:By: Dotan Persitz \n Advisors: \n Where: Bloomfield 424 From:
\nTel-Aviv University.\nAbstract: \n\n \;\n\nWe study dynamic task al
location when there is a fixed bipartite network associating workers to ta
sks. We analyze two approaches: centralized and decentralized. First\, we
study the optimal policy of a planner whose objective is to minimize the e
xpected time of completion of all tasks. Second\, we analyse a game played
by workers who independently choose their tasks and are rewarded each tim
e they complete a task. We characterize networks for which the planner's a
nd workers' policies are simple.\n\nWhen policies are simple the planner p
refers the workers to start with the hardest tasks\, whereas workers alway
s prefer to start with easier tasks. We show that the two policies only co
incide when the bipartite network satisfies a strong symmetry condition. D
ifferential rewards can be used to implement the planner's optimal task al
location and we show that non-contingent rewards\, which are independent o
f the set of remaining tasks\, can be used as long as there is no task tha
t a single agent can complete.\n\n \;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:430@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240214T113000
DTEND;TZID=Asia/Jerusalem:20240214T123000
DTSTAMP:20240213T141439Z
URL:https://dds.technion.ac.il/ar/iemevents/centrality-measure-for-opinion
-dynamics/
SUMMARY:Centrality Measure for Opinion Dynamics [ \n Game Theory Semina
r\n Seminars\n \n ]
DESCRIPTION:By: Ron Peretz \n Advisors: \n Where: Bloomfield 424 From:\nB
ar-Ilan University.\nAbstract: \n We propose a centrality measure for op
inion dynamics that utilizes the game-theoretic concept of “authority di
stribution” due to [Hu\, Shapley 2003]. An axiomatic foundation is given
in the tradition of coalitional game theory. Several extensions are devel
oped. A novelty of our approach is that it is based on the opinion exchang
e dynamics rather than network topology. However\, for any fixed exchange
dynamics\, our approach associates a network centrality measure with any g
iven network topology. Previous as well as new network centrality measures
are obtained in this way.\n \n \nJoint work with Igal Milchtaich and Y
evgeny Tsodikovich
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:432@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240221T111500
DTEND;TZID=Asia/Jerusalem:20240221T121500
DTSTAMP:20240221T061204Z
URL:https://dds.technion.ac.il/ar/iemevents/actual-causality-a-survey/
SUMMARY:Actual Causality: A Survey [ \n Game Theory Seminar\n Semin
ars\n \n ]
DESCRIPTION:By: Joe Halpern \n Advisors: \n Where: Bloomfield 424 From:\n
Cornell University \nAbstract: \n\nWhat does it mean that an event C ``ac
tually caused'' event E? The problem of defining actual causation goes bey
ond mere philosophical\n\nspeculation. For example\, in many legal argum
ents\, it is precisely what needs to be established in order to determine
responsibility. (What exactly was the actual cause of the car accident or
the medical problem?) The philosophy literature has been struggling with t
he problem\n\nof defining causality since the days of Hume\, in the 1700s.
Many of the definitions have been couched in terms of counterfactuals.\n\
n(C is a cause of E if\, had C not happened\, then E would not have happen
ed.) In 2001\, Judea Pearl and I introduced a new definition of actual cau
se\, using Pearl's notion of structural equations to model counterfactuals
. The definition has been revised twice since then\,\n\nextended to deal
with notions like "responsibility" and "blame"\, and applied in databases
and program verification.\n\nI survey the last 15 years of work here\, in
cluding joint work with Judea Pearl\, Hana Chockler\, and Chris Hitchcock.
\n\nThe talk will be completely self-contained.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:439@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240228T113000
DTEND;TZID=Asia/Jerusalem:20240228T123000
DTSTAMP:20240228T061833Z
URL:https://dds.technion.ac.il/ar/iemevents/common-knowledge-regained/
SUMMARY:Common Knowledge Regained [ \n Game Theory Seminar\n Semina
rs\n \n ]
DESCRIPTION:By: Yoram Moses \n Advisors: \n Where: Bloomfield 424 From:\n
Technion\nAbstract: \nOver the last six decades\, common knowledge has as
sumed a central role in assumptions underlying the study of economic inter
actions\, and in the design and analysis of multi-agent systems. Somewhat
paradoxically\, even small timing frictions\, which are very common in pra
ctice\, make attaining common knowledge of new facts impossible. This unin
tuitive phenomenon\, discussed among others by Arrow et al. in 1987 and b
y Aumann in 1989\, has evaded satisfactory resolution for four decades. Th
is talk will present and discuss a new definition of common knowledge that
resolves the paradox. This definition coincides with the traditional defi
nition in static systems\, generalizes it\, and overcomes some of its shor
tcomings in dynamic settings. The talk will be self contained and will not
assume prior familiarity with the topic.\n \;\nJoint work with Yanna
i Gonczarowski
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:440@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240228T130000
DTEND;TZID=Asia/Jerusalem:20240228T140000
DTSTAMP:20240226T105827Z
URL:https://dds.technion.ac.il/ar/iemevents/network-formation-for-learning
/
SUMMARY:Network Formation for Learning [ \n Game Theory Seminar\n S
eminars\n \n ]
DESCRIPTION:By: Lior Sasson \n Advisors: \n Where: Bloomfield 527 From:\n
Technion\nAbstract: \n\nWe study a network formation interaction in which
the formatted network serves the agents to form beliefs about an unknown
state via the DeGroot naive learning process. Consequently\, the value of
the network captures the ability of the network to aggregate information.
We focus on the notion of stable networks in which agents do not want to c
reate or erase any of their edges. Our main result provides structural pro
perties of stable networks. We show that in every stable network\, the deg
ree of any two vertices differs by at most one. As a corollary we get that
any stable network aggregates information well and a price of anarchy upp
er bound is provided.\nJoint work with Itai Arieli and Yakov Babichenko
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:442@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240306T113000
DTEND;TZID=Asia/Jerusalem:20240306T123000
DTSTAMP:20240304T082653Z
URL:https://dds.technion.ac.il/ar/iemevents/liquidity-games-a-new-type-of-
repeated-bayesian-games-to-model-financial-capital-markets/
SUMMARY:Liquidity Games: A new type of repeated Bayesian games to model fin
ancial (capital) markets. [ \n Game Theory Seminar\n Seminars\n
\n ]
DESCRIPTION:By: Alicia Vidler \n Advisors: \n Where: Bloomfield 424 From:
\nUniversity of New South Wales\nAbstract: \n \nWe present a new type of g
ame: the Liquidity Game. We draw inspiration from the UK government bond m
arket and apply game theoretic approaches to its analysis. In Liquidity Ga
mes\, market participants (agents) use non-cooperative games where the pla
yers' utility is directly defined by the liquidity of the game itself\, of
fering a paradigm shift in our understanding of market dynamics. Each pla
yer’s utility is intricately linked to the liquidity generated within th
e game\, making the utility endogenous and dynamic. Players are not just p
assive recipients of utility based on external factors but active particip
ants whose strategies and actions collectively shape and are shaped by the
liquidity of the market. This reflexivity introduces a level of complexit
y and realism previously un-attained in conventional models. We apply Liq
uidity Game theoretic approaches to a simple UK bond market interaction an
d present results for market design and strategic behaviour of participant
s. My presentation assumes no prior knowledge of financial markets by th
e audience. To help discussion\, I will present a short description of ho
w the bond markets in question currently work. \n
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:444@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240313T113000
DTEND;TZID=Asia/Jerusalem:20240313T123000
DTSTAMP:20240311T053713Z
URL:https://dds.technion.ac.il/ar/iemevents/minimal-contagious-sets-degree
-distributional-bounds/
SUMMARY:Minimal Contagious Sets: Degree Distributional Bounds [ \n Game
Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Yevgeny Tsodikovich \n Advisors: \n Where: Bloomfield 424
From:\nBar-Ilan University\nAbstract: \n\nWe study the minimum seed si
ze necessary for successful innovation adoption in networks of growing siz
e\, such as scale-free networks\, widely used to model real-world social n
etworks. We employ reply dynamics in a 2x2 coordination game\, where one a
ction represents innovation\, and the other represents the status quo. Age
nts adopt the innovation if a certain fraction of their neighbors have alr
eady adopted it. Our main contribution is to provide upper and lower bound
s on the seed set's size and to demonstrate that in growing networks\, the
minimum seed required to reach the entire network cannot be sub-linear
in the network size. Our results establish a theoretical foundation for th
e optimal seeding strategies within more complex network structures and dy
namics.\n\n \;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:449@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240320T113000
DTEND;TZID=Asia/Jerusalem:20240320T123000
DTSTAMP:20240318T125402Z
URL:https://dds.technion.ac.il/ar/iemevents/naive-calibration/
SUMMARY:Naive Calibration [ \n Game Theory Seminar\n Seminars\n \
n ]
DESCRIPTION:By: Benjamin Bachi\n Advisors: \n Where: Bloomfield 424 From:
\nHaifa University\nAbstract: \n We develop a model of non-Bayesian deci
sion-making in which an agent obtains an estimate of the state of a releva
nt economic fundamental but does not know the joint distribution of the tw
o. To make use of the estimate\, she relies on an endogenously generated d
ataset that consists of previous estimates and state realizations. She att
ributes a systematic difference between the estimates and state realizatio
ns in her dataset to a systematic bias in the estimate and naively calibra
tes it. Her subsequent action affects the probability with which the estim
ate and the corresponding state realization will be recorded in the datase
t that will be used in future decisions. We investigate the steady state o
f the naive calibration procedure and show that it results in a seemingly
pessimistic behavior that is exacerbated by feedback loops. We apply our m
odel to project selection problems and second-price IPV auctions.\n
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:451@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240327T113000
DTEND;TZID=Asia/Jerusalem:20240327T123000
DTSTAMP:20240325T064338Z
URL:https://dds.technion.ac.il/ar/iemevents/investment-decisions-voluntary
-disclosure-myopia-and-bounded-inefficiency/
SUMMARY:Investment Decisions\, Voluntary Disclosure\, Myopia\, and Bounded
Inefficiency [ \n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Ilan Kremer \n Advisors: \n Where: Bloomfield 424 From:\n
Hebrew University\nAbstract:\n\nWe present a strategic disclosure model wh
ere a manager chooses between a short-term and a long-term project. The sh
ort-term cash flow can be disclosed early but can also be temporarily hidd
en. Without disclosure\, the stock price reflects suspicions that the mana
ger is concealing a negative outcome of the short-term project. Consistent
with the existing literature\, we find that strategic disclosure may lead
to inefficiency\, with the manager selecting an inferior short-term proje
ct to maximize the short-run price. However\, our main result demonstrates
that the inefficiency is quite limited\, even when the manager cares almo
st entirely about the short-term stock price.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:468@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240605T113000
DTEND;TZID=Asia/Jerusalem:20240605T123000
DTSTAMP:20240530T052655Z
URL:https://dds.technion.ac.il/ar/iemevents/robust-contracting-and-volunta
ry-disclosure/
SUMMARY:Robust contracting and voluntary disclosure [ \n Game Theory Se
minar\n Seminars\n \n ]
DESCRIPTION:By: Eran Hanany \n Advisors: \n Where: Bloomfield 424 From:\nT
el-Aviv University \nAbstract:\n\n \;\n\nThis paper analyzes contracti
ng between a principal and an agent when the principal is uncertain exactl
y which actions may be feasible for the agent and has a strong desire for
robustness (in the worst-case or maxmin sense) of the expected profits gen
erated. A prominent and path-breaking paper in this direction is Carroll (
2015)\, which demonstrates that linear contracts are robustly (worst-case)
optimal given uncertainty about an agent’s available actions. What if\,
when it is in their interest\, the agent could choose to disclose that th
ey have access to a particular additional action\, and such statements cou
ld be verified by the principal? Does this change the form of robustly opt
imal contracts offered to an agent who either chooses not to disclose or h
as no additional action to disclose? Are such contracts still linear? We s
how that voluntary disclosure can substantially change the form of robustl
y optimal contracts. In particular\, we show the possibility of and provid
e sufficient conditions for equilibrium contracts offered following non-di
sclosure to be non-linear. This equilibrium non-linearity does not always
occur. We show that linearity results when there are few publicly known-to
-be-available actions that generate a positive surplus.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:469@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240605T130000
DTEND;TZID=Asia/Jerusalem:20240605T140000
DTSTAMP:20240530T052840Z
URL:https://dds.technion.ac.il/ar/iemevents/continuous-thieles-rules-for-d
istribution-aggregation/
SUMMARY:Continuous Thiele's Rules for Distribution Aggregation. [ \n Ga
me Theory Seminar\n Seminar in Probability and Stochastic Processes\n
Seminars\n \n ]
DESCRIPTION:By: Jonathan Wagner \n Advisors: \n Where: Bloomfield 424 From
:\nTechnion\nAbstract: \n\nWe introduce the class of Continuous Thiele’
s Rules that generalize the familiar Thiele’s rules of multi-winner voti
ng to distribution aggregation problems. Each rule in that class maximizes
the sum of f(u) over all agents where u is utility and f could be any dif
ferentiable\, increasing\, and concave real function. We show how bounds
on Egalitarian loss\, welfare loss\, and the approximation rate of Averag
e Justified Representation can be derived from the "Relative Risk Aversi
on" of f\, leading to a quantifiable\, continuous presentation of their in
evitable trade-offs.\n\n \;\n\n \;
CATEGORIES:Game Theory Seminar,Seminar in Probability and Stochastic
Processes,Seminars
END:VEVENT
BEGIN:VEVENT
UID:482@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240626T113000
DTEND;TZID=Asia/Jerusalem:20240626T123000
DTSTAMP:20240625T081139Z
URL:https://dds.technion.ac.il/ar/iemevents/vaguely-good-news-and-accurate
-bad-news-a-modular-approach/
SUMMARY:Vaguely Good News and Accurate Bad News: A Modular Approach [ \n
Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Amnon Schreiber \n Advisors: \n Where: Bloomfield 424 Fro
m:\nBar-Ilan University\nAbstract: \n\nWe analyze a model where a firm vo
luntarily discloses verifiable signals about its value. A Bayesian market\
, unaware of the specific signals the firm holds\, sets the firm price bas
ed on Bayes’ rule. Our main result indicates that the firm tends to disc
lose "bad signals" only if they are accurate. Conversely\, "good signals"
will be disclosed even if they are noisy. We focus on a specific equilibri
um refinement\, which we call "Modular." This modular equilibrium exhibits
many interesting and beneficial features\, which we discuss in detail.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:486@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240703T113000
DTEND;TZID=Asia/Jerusalem:20240703T123000
DTSTAMP:20240701T085132Z
URL:https://dds.technion.ac.il/ar/iemevents/heterogeneous-noise-and-stable
-miscoordination/
SUMMARY:Heterogeneous Noise and Stable Miscoordination [ \n Game Theory
Seminar\n Seminars\n \n ]
DESCRIPTION:By: Yuval Heller \n Advisors: \n Where: Bloomfield 424 From:\
nBar-Ilan University\nAbstract: \n\nCoordination games admit two types of
equilibria: coordinated pure equilibria in which everyone plays the same
action\, and inefficient mixed equilibria with miscoordination. The existi
ng literature shows that populations will converge to one of the pure coor
dinated equilibria from almost any initial state. By contrast\, we show th
at plausible learning dynamics\, in which agents sample the aggregate beha
vior of the opponent’s population and best reply to their samples\, can
induce stable miscoordination if there is heterogeneity in the sample size
s: some agents base their choices on noisy small samples (anecdotal eviden
ce)\, while others rely on large samples.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:489@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240710T113000
DTEND;TZID=Asia/Jerusalem:20240710T123000
DTSTAMP:20240709T052103Z
URL:https://dds.technion.ac.il/ar/iemevents/persuading-while-learning/
SUMMARY:Persuading While Learning [ \n Game Theory Seminar\n Semina
rs\n \n ]
DESCRIPTION:By: Dima Shaiderman \n Advisors: \n Where: Bloomfield 424 From
:\nTechnion\nAbstract: \n\nWe propose a dynamic product adoption persuasi
on model involving an impatient partially informed sender who gradually
learns the state. In this model\, the sender gathers information over t
ime\, and hence her posteriors' sequence forms a discrete-time martingal
e. The sender commits to a dynamic revelation policy to persuade the agen
t to adopt a product. We demonstrate that under the assumption that the s
ender's martingale is Markovian and possesses Blackwell-preserving kernels
\, the family of optimal strategies for the sender takes an interval form\
; namely\, in every period the set of martingale realizations in which a
doption occurs is an interval. Utilizing this\, we prove that if the sen
der is sufficiently impatient\, then under a random walk martingale\, the
optimal policy is fully transparent up to the moment of adoption\; namel
y\, the sender reveals the entire information she privately holds in every
period.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:491@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240724T113000
DTEND;TZID=Asia/Jerusalem:20240724T123000
DTSTAMP:20240724T060825Z
URL:https://dds.technion.ac.il/ar/iemevents/envious-explore-and-exploit-2/
SUMMARY:Envious Explore and Exploit [ \n Game Theory Seminar\n Semi
nars\n \n ]
DESCRIPTION:By: Omer Ben-Porat \n Advisors: \n Where: Bloomfield 424 From:
\nTechnion\nBased on a joint project with Yotam Gafni and Or Markovetzki\n
\nAbstract: \n\nExploration and exploitation play key roles in recommenda
tion systems (RSs)\, aiming to better serve users by learning from the env
ironment. Despite their success in the past decade\, the societal effects
of RSs are not well understood\, particularly regarding the utility discre
pancy between different users. In this work\, we quantify such discrepanci
es using the economic notion of envy. We present a Bayesian Multi-armed Ba
ndits model in which each round consists of several sessions\, with reward
s realized once per round. We refer to this property as reward consistency
and demonstrate that the RS can leverage it for better societal outcomes.
However\, this approach also generates envy\, as late-arriving users bene
fit from the information gathered by early-arriving users. We examine the
envy generated under several arrival order mechanisms. We provide tight en
vy bounds for a large class of RSs with uniform arrival and upper bound th
e envy for controlled arrival\, where the RS can influence the order of ar
rival by nudging its users. Furthermore\, we study the efficiency-fairness
trade-off by devising an algorithm that maintains constant envy and appro
ximates optimal welfare in restricted settings.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:498@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240731T113000
DTEND;TZID=Asia/Jerusalem:20240731T123000
DTSTAMP:20240729T053333Z
URL:https://dds.technion.ac.il/ar/iemevents/clerks/
SUMMARY:Clerks [ \n Game Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Kfir Eliaz \n Advisors: \n Where: Bloomfield 424 From:\nT
el-Aviv University\nAbstract: \n\nWe study the optimal dynamic scheduling
of workers to tasks when task-completion is privately observed (hence\, w
orkers can delay the release of completed tasks)\, and when idle time is t
he only means of providing incentives. Our main result characterizes a sch
eduling rule\, and the equilibrium it induces\, maximizing the expected di
scounted output subject to workers' incentive constraints. When workers ar
e inherently slow\, a simple rotation scheme suffices to attain first-best
output\, but when they are more productive\, optimal scheduling alternate
s between phases with and without delay. Our analysis highlights a trade-o
ff between the quality and size of workforce.\n\n \;\n\nBased on a joi
nt work with Daniel Fershtman and Alexander Frug\n\nA link to the paper.\n
\n \;\n\n \;
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VEVENT
UID:501@dds.technion.ac.il
DTSTART;TZID=Asia/Jerusalem:20240807T113000
DTEND;TZID=Asia/Jerusalem:20240807T123000
DTSTAMP:20240806T043334Z
URL:https://dds.technion.ac.il/ar/iemevents/demand-for-online-news-inertia
-and-misperceptions/
SUMMARY:Demand for Online News\, Inertia\, and Misperceptions [ \n Game
Theory Seminar\n Seminars\n \n ]
DESCRIPTION:By: Roee Levy \n Advisors: \n Where: Bloomfield 424 From:\nTe
l-Aviv University\nAbstract: \n\nThere has been growing concern that indi
viduals consume low-quality\, extreme\, news on social media. We test whet
her such news choices may stem from misperceptions and whether simple inte
rventions can decrease the slant and increase the quality of people's news
diets on social media. In a survey experiment\, we first elicit perceptio
ns regarding the slant and quality of news and then randomize participants
into three groups: a group provided with information on media slant\, a g
roup provided with information on media quality\, and a group not provided
with any information. After the intervention\, we offer all participants
an option to readjust their news portfolio by following or unfollowing new
s outlets on Facebook. We find that individuals suffer from misperceptions
\, especially regarding the quality of outlets. Correcting these mispercep
tions increases the quality of the outlets individuals choose to follow. H
owever\, providing information on slant does not have a significant effect
on news choices. Still\, we find that simply providing individuals an opt
ion to readjust their news choices results in a more moderate news portfol
io.
CATEGORIES:Game Theory Seminar,Seminars
END:VEVENT
BEGIN:VTIMEZONE
TZID:Asia/Jerusalem
X-LIC-LOCATION:Asia/Jerusalem
BEGIN:STANDARD
DTSTART:20201025T010000
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:IST
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20210326T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0300
TZNAME:IDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20211031T010000
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:IST
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20220325T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0300
TZNAME:IDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20221030T010000
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:IST
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20230324T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0300
TZNAME:IDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20231029T010000
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:IST
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20240329T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0300
TZNAME:IDT
END:DAYLIGHT
END:VTIMEZONE
END:VCALENDAR