#### DMCA

## Computer science and game theory (2008)

### Cached

### Download Links

- [ai.stanford.edu]
- [robotics.stanford.edu]
- [ai.stanford.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | Communications of the ACM |

Citations: | 18 - 1 self |

### Citations

2438 |
The Foundations of Statistics
- Savage
- 1956
(Show Context)
Citation Context ...e key to achieving von Neumann and Morgenstern’s third stage, and may prove to be a joint endeavor between computer science and game theory. 5 Parenthetically it can be remarked that Savage’s setting =-=[34]-=- on which the modern Bayesian framework is based does not have an obvious extension to the multi-agent case. However this is not the focus of the point I’m making here. 7Acknowledgements I thank Alon... |

1838 |
Reasoning about Knowledge
- FAGIN, HALPERN, et al.
- 1995
(Show Context)
Citation Context ...on between computer science (including distributed systems, AI and theory) and game theory. Beside game theory, it made deep ties with philosophy and mathematical logic, and culminated in the seminal =-=[8]-=-. 4 It is interesting to speculate why this area was omitted from Kalai’s list even although it predates his paper by a decade, and why today it is not as active as the other areas. I think this is be... |

1015 |
Intention is choice with commitment
- Cohen, Levesque
- 1990
(Show Context)
Citation Context ...h effect intelligent behavior and reason about it. Philosophers have written about it (e.g., [1]), and there have been attempts – albeit preliminary ones – to formalize these intuition (starting with =-=[3]-=-). Some in AI have advocated embracing an even broader vocabulary of emotions (e.g., the recent provocative albeit informal [23]). Is game theory missing out by not considering these concepts? 4 Concl... |

656 | Algorithmic mechanism design, in
- Nisan, Ronen
- 1999
(Show Context)
Citation Context ...incentive compatibility, individual rationality, and social welfare maximization. Good examples include the interaction between the Vickrey-Clarke-Groves (VCG) mechanism and shortest-path computation =-=[26]-=-, and the literature on combinatorial auctions [6] which combines a weighted-set-packing-like NP-hard optimization problem with incentive issues. In this connection one might also mention the interpla... |

571 | Algorithmic Game Theory
- Nisan, Roughgarden, et al.
- 2007
(Show Context)
Citation Context ...ns, information markets, and reputation systems. This combination of third and fourth areas is arguably the most active area today at the interface of CS and GT, and many aspects of it are covered in =-=[25]-=-, an extensive edited collection of surveys. The popularity of this area is perhaps not surprising. The relevance of specific applications speak for itself (although arguments remain about whether the... |

494 |
Lectures on Modeling Bounded Rationality
- RUBINSTEIN
- 1996
(Show Context)
Citation Context ...ve modelling of agents. Backing off from these strong assumptions has proven challenging. A fairly thin strand of work under the heading of “bounded rationality” studies games played by automata (cf. =-=[33]-=-). This is an important area of research, and sometimes makes deep connections between the two fields. Early results, for example, showed that one of the well known pesky facts in game theory – namely... |

443 | Combinatorial Auctions
- Cramton, Shoham, et al.
- 2005
(Show Context)
Citation Context ...d social welfare maximization. Good examples include the interaction between the Vickrey-Clarke-Groves (VCG) mechanism and shortest-path computation [26], and the literature on combinatorial auctions =-=[6]-=- which combines a weighted-set-packing-like NP-hard optimization problem with incentive issues. In this connection one might also mention the interplay between mechanism design and cryptography. They ... |

280 | Graphical models for game theory
- Kearns, Littman, et al.
- 2001
(Show Context)
Citation Context ...been picked up as vigorously as they might. There has been substantial work on compact and otherwise specialized game representations. Some of them are indeed graph-based, for example graphical games =-=[18]-=-, local-effect games [21], MAIDS [19], and Game networks [20]. The graph-based representations extend also to coalition game theory [7]. But specialized representations exist that are not graph based,... |

208 | Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations
- Shoham, Leyton-Brown
- 2009
(Show Context)
Citation Context ...on relevant concepts from complexity theory. The material discussed so far features prominently in computer science conferences and journals, and is also beginning to find its way into textbooks (cf. =-=[25, 35]-=-). These areas will undoubtedly continue to flourish, but I want to turn the attention to the two closely related areas – 4 and 5 – listed by Kalai that have not been looked as closely by the communit... |

198 | On the complexity of the parity argument and other inefficient proofs of existence
- Papadimitriou
- 1994
(Show Context)
Citation Context ...cepts) has indeed been the focus of much interest in CS, especially within the theory community. A new complexity class – PPAD – was proposed to handle problems for which a solution is known to exist =-=[27]-=-, the computation of a sample Nash equilibrium was shown to be complete for this class [2], and the problem of computing Nash equilibria with specific properties was shown to be NP-hard [10, 4]. At th... |

185 | Multi-agent influence diagrams for representing and solving games
- Koller, B
- 2002
(Show Context)
Citation Context ...might. There has been substantial work on compact and otherwise specialized game representations. Some of them are indeed graph-based, for example graphical games [18], local-effect games [21], MAIDS =-=[19]-=-, and Game networks [20]. The graph-based representations extend also to coalition game theory [7]. But specialized representations exist that are not graph based, for example multi-attribute based [5... |

168 |
On the complexity of cooperative solution concepts
- Deng, Papadimitriou
- 1994
(Show Context)
Citation Context ...ome of them are indeed graph-based, for example graphical games [18], local-effect games [21], MAIDS [19], and Game networks [20]. The graph-based representations extend also to coalition game theory =-=[7]-=-. But specialized representations exist that are not graph based, for example multi-attribute based [5] and logic based [15]. I believe this area is ripe for additional work, for example one that has ... |

150 |
Nash and correlated equilibria: Some complexity considerations
- Gilboa, Zemel
- 1989
(Show Context)
Citation Context ... to exist [27], the computation of a sample Nash equilibrium was shown to be complete for this class [2], and the problem of computing Nash equilibria with specific properties was shown to be NP-hard =-=[10, 4]-=-. At the same time, algorithms – some quite sophisticated, and all exponential in the worst case – have been proposed to compute Nash equilibria [41, 11]. Somewhat surprisingly, it was recently shown ... |

133 | Complexity results about nash equilibria
- Conitzer, Sandholm
- 2003
(Show Context)
Citation Context ... to exist [27], the computation of a sample Nash equilibrium was shown to be complete for this class [2], and the problem of computing Nash equilibria with specific properties was shown to be NP-hard =-=[10, 4]-=-. At the same time, algorithms – some quite sophisticated, and all exponential in the worst case – have been proposed to compute Nash equilibria [41, 11]. Somewhat surprisingly, it was recently shown ... |

114 |
Settling the complexity of 2-player nash-equilibrium
- Chen, Deng
- 2006
(Show Context)
Citation Context ...nity. A new complexity class – PPAD – was proposed to handle problems for which a solution is known to exist [27], the computation of a sample Nash equilibrium was shown to be complete for this class =-=[2]-=-, and the problem of computing Nash equilibria with specific properties was shown to be NP-hard [10, 4]. At the same time, algorithms – some quite sophisticated, and all exponential in the worst case ... |

110 | Simple search methods for finding a nash equilibrium
- Porter, Nudelman, et al.
- 2004
(Show Context)
Citation Context ...osed to compute Nash equilibria [41, 11]. Somewhat surprisingly, it was recently shown experimentally that a relatively simple search algorithm significantly outperforms more sophisticated algorithms =-=[31]-=-. This is a very active area that promises many additional results. The third match is somewhat less tight than the first two. There are at least two kinds of optimization one could speak about in a g... |

72 | Computing Shapley values, manipulating value division schemes, and checking core membership in multi-issue domains. AAAI
- Conitzer, Sandholm
- 2004
(Show Context)
Citation Context ...9], and Game networks [20]. The graph-based representations extend also to coalition game theory [7]. But specialized representations exist that are not graph based, for example multi-attribute based =-=[5]-=- and logic based [15]. I believe this area is ripe for additional work, for example one that has the strategy space of agents described using constructs of programming languages. The complexity of com... |

59 |
A global newton method to compute nash equilibria
- Govindan, Wilson
- 2003
(Show Context)
Citation Context ...ith specific properties was shown to be NP-hard [10, 4]. At the same time, algorithms – some quite sophisticated, and all exponential in the worst case – have been proposed to compute Nash equilibria =-=[41, 11]-=-. Somewhat surprisingly, it was recently shown experimentally that a relatively simple search algorithm significantly outperforms more sophisticated algorithms [31]. This is a very active area that pr... |

59 | Computing equilibria for two-person games
- Stengel
- 2002
(Show Context)
Citation Context ...ith specific properties was shown to be NP-hard [10, 4]. At the same time, algorithms – some quite sophisticated, and all exponential in the worst case – have been proposed to compute Nash equilibria =-=[41, 11]-=-. Somewhat surprisingly, it was recently shown experimentally that a relatively simple search algorithm significantly outperforms more sophisticated algorithms [31]. This is a very active area that pr... |

52 | Local-effect games
- Leyton-Brown, Tennenholtz
- 2003
(Show Context)
Citation Context ...sly as they might. There has been substantial work on compact and otherwise specialized game representations. Some of them are indeed graph-based, for example graphical games [18], local-effect games =-=[21]-=-, MAIDS [19], and Game networks [20]. The graph-based representations extend also to coalition game theory [7]. But specialized representations exist that are not graph based, for example multi-attrib... |

51 |
Autonomous Bidding Agents: Strategies and Lessons from the Trading Agent Competition
- Wellman, Greenwald, et al.
- 2007
(Show Context)
Citation Context ...eresy, but real life suggests that perhaps we have no choice. For example, in the trading agent competition (TAC), Nash equilibrium of the game did not play a role in almost any participating program =-=[42]-=-, and this is certainly true of the more established chess and checkers competitions. It is premature to write off the Nash equilibrium as irrelevant. For example, one program (see chapter 8 of [42]) ... |

34 |
The Emotion Machine: Commonsense Thinking, Artificial Intelligence, and the Future of the Human Mind
- Minsky
(Show Context)
Citation Context ... – albeit preliminary ones – to formalize these intuition (starting with [3]). Some in AI have advocated embracing an even broader vocabulary of emotions (e.g., the recent provocative albeit informal =-=[23]-=-). Is game theory missing out by not considering these concepts? 4 Concluding Remarks Science operates at many levels. For some, it is sufficient that scientific theories be clever, beautiful and insp... |

33 | Nonmonotonic reasoning
- Brewka, Niemelä, et al.
- 2008
(Show Context)
Citation Context ...hereas in computer science – in particular, in database theory and artificial intelligence – belief revision and other dynamic theories [30] (including the entire mini-industry of nonmonotonic logics =-=[9]-=-) play an equal if not greater role. Indeed, recent work at the interface of logic and game theory [37] extends the static treatment of [8] in a dynamic direction. 4to the following additional survey... |

28 | Program equilibrium
- Tennenholtz
(Show Context)
Citation Context ... finite automata with sufficiently few states [24, 28]. A more recent result shows that when players in a game are computer programs, one obtains phenomena akin to the Folk Theorem for repeated games =-=[36]-=-. This connection between theoretical models of computation and game theory is I believe quite important and beautiful, but a fairly narrow interpretation of the term ‘bounded rationality’. The term s... |

26 |
Theory of Games and Economic Behavior, Second edition
- Neumann, Morgenstern
- 1947
(Show Context)
Citation Context ...ach other mutually. Beyond this lies the field of real success: genuine predictions by theory. It is well known that all mathematized sciences have gone through these successive phases of evolution. (=-=[40]-=-, pp. 7-8) So at least von Neumann, the father of modern-day game theory and computer science, attached importance to spanning the spectrum from the theoretical to the applied. Pragmatics may be key t... |

21 | A computer scientist looks at game theory
- Halpern
(Show Context)
Citation Context ...ogic, and others. • [29], geared towards computer scientists, is a concise, 5-page paper summarizing the main complexity and algorithmic issues at the interface of CS and GT circa 2001. • The 21-page =-=[13]-=- is similar to [22] in that it is geared towards game theorists and its main focus is distributed systems, but it is more current than the latter given the decade separating the two. [13] later evolve... |

19 | Computer science and game theory: A brief survey
- Halpern
- 2008
(Show Context)
Citation Context ...ilar to [22] in that it is geared towards game theorists and its main focus is distributed systems, but it is more current than the latter given the decade separating the two. [13] later evolved into =-=[14]-=-, a 17-page survey with an abbreviated discussion of distributed computing, but with additional material on complexity considerations, price of anarchy, mediators, and other topics. • The 30-page [32]... |

19 |
Multi-attribute coalition games
- Ieong, Shoham
- 2006
(Show Context)
Citation Context ...s [20]. The graph-based representations extend also to coalition game theory [7]. But specialized representations exist that are not graph based, for example multi-attribute based [5] and logic based =-=[15]-=-. I believe this area is ripe for additional work, for example one that has the strategy space of agents described using constructs of programming languages. The complexity of computing a sample Nash ... |

17 | A new algorithm for generating equilibria in massive zero-sum games
- Zinkevich, Bowling, et al.
- 2007
(Show Context)
Citation Context ...in fact make use of what can be viewed as approximate empirical NE. Another striking example is the computation of equilibria in a simplified game tree by a top scoring program in a poker competition =-=[43]-=-. It could be argued that maxmin strategies, which coincide with equilibrium strategies in zero-sum games, do play an important pragmatic role. But certainly computation of either maxmin or equilibriu... |

16 |
Game-theoretical aspects of computing
- Linial
- 1994
(Show Context)
Citation Context ...itional surveys, all by computer scientists, each with a slightly different slant, and most of which go in considerably more detail into some of the topics. • The earliest relevant survey is probably =-=[22]-=-. Geared primarily towards game theorists, this 58-page survey has deep coverage of game theoretic aspects of distributed systems, fault-tolerant computing, and cryptography, and also touches on other... |

12 |
Game networks
- La-Mura
- 2000
(Show Context)
Citation Context ...bstantial work on compact and otherwise specialized game representations. Some of them are indeed graph-based, for example graphical games [18], local-effect games [21], MAIDS [19], and Game networks =-=[20]-=-. The graph-based representations extend also to coalition game theory [7]. But specialized representations exist that are not graph based, for example multi-attribute based [5] and logic based [15]. ... |

12 | Computing equilibria: a computational complexity perspective. Econ Theory 42
- Roughgarden
- 2010
(Show Context)
Citation Context ...[14], a 17-page survey with an abbreviated discussion of distributed computing, but with additional material on complexity considerations, price of anarchy, mediators, and other topics. • The 30-page =-=[32]-=- is a detailed survey of a specific topic, namely the complexity of computing a sample Nash equilibrium. Geared mostly towards economists, it includes ample background material on relevant concepts fr... |

8 |
Exploring logical dynamics. Center for the Study of Language and Information
- Benthem
- 1997
(Show Context)
Citation Context ...vision and other dynamic theories [30] (including the entire mini-industry of nonmonotonic logics [9]) play an equal if not greater role. Indeed, recent work at the interface of logic and game theory =-=[37]-=- extends the static treatment of [8] in a dynamic direction. 4to the following additional surveys, all by computer scientists, each with a slightly different slant, and most of which go in considerab... |

5 | On bounded rationality and computational complexity
- Papadimitriou, Yannakakis
- 1994
(Show Context)
Citation Context ...that constant ‘defection’ is the only subgame-perfect equilibrium in the finitelyrepeated prisoner’s dilemma game – ceases to hold true if the players are finite automata with sufficiently few states =-=[24, 28]-=-. A more recent result shows that when players in a game are computer programs, one obtains phenomena akin to the Folk Theorem for repeated games [36]. This connection between theoretical models of co... |

3 |
Introduction to the special issue on learning and computational game theory
- Greenwald, Littman
- 2007
(Show Context)
Citation Context ...discovered with something of a vengeance in computer science and in particular AI, witness special issues devoted to it in the Journal of Artificial Intelligence [39] and the Machine Learning Journal =-=[12]-=-. For computer science, the move from single-agent learning to multiagent learning is interesting not only because it calls for new solutions, but because the very questions change. When multiple agen... |

3 | When are two games the same
- Benthem
- 1998
(Show Context)
Citation Context ...on. It is perhaps presumptuous for CS to intrude on this debate, except insofar as it lends logical insights, since I do think that logic is a useful lens through which to look at 5these issues (cf. =-=[38]-=-). But perhaps indeed this is more the role of mathematical logic than of CS per se. Where CS can truly lead the way is I think on the pragmatics of game theory. Game theory as we know it embodies rad... |

2 |
Presidential address, the second World Congress of the Game Theory
- Kalai
- 2008
(Show Context)
Citation Context ...ubjected to empirical evaluation. Without personally weighing in on this emotional debate, I note that in his presidential address at the quad-annual congress of the International Game Theory Society =-=[17]-=-, Kalai reprised the three stages of any science as discussed by von Neumann and Morgenstern: [W]hat is important is the gradual development of a theory, based on a careful analysis of the ordinary ev... |

2 |
Belief revision
- Pappas
- 2007
(Show Context)
Citation Context .... In game theory static theories are indeed the primary focus, whereas in computer science – in particular, in database theory and artificial intelligence – belief revision and other dynamic theories =-=[30]-=- (including the entire mini-industry of nonmonotonic logics [9]) play an equal if not greater role. Indeed, recent work at the interface of logic and game theory [37] extends the static treatment of [... |

1 |
Wellman (eds.). Special issue on foundations of multiagent learning
- Vohra, P
(Show Context)
Citation Context ...or focus within game theory, has been discovered with something of a vengeance in computer science and in particular AI, witness special issues devoted to it in the Journal of Artificial Intelligence =-=[39]-=- and the Machine Learning Journal [12]. For computer science, the move from single-agent learning to multiagent learning is interesting not only because it calls for new solutions, but because the ver... |