D. Blackwell and M. A. Girshnik. Theory of Games and Statistical Decisions. Dover Publications, New York, NY, 1979.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....investigate three gray box Information and Control Layers (ICLs) or determining the contents of the file cache, controlling he layout of files across local disk, and limiting process ezecution based on available memory. A gray box ICL sits between a client and the OS and uses a combination of a19orithmic knowledge, observations, and in]erences to garner information about or control the behavior of a gray boa system. We summarize a set of techniques that are helpful in building gray box [CLs and have begn to organizq a gray toolbox to ease the con struction of ICLs. Through our case ....

....system; we believe that this knowledge should be encapsulated in ICLs, so that these tecliniques can be used by all programmers. However: gray box systems go one step further by combining knowledge with measurements and observations, a technique com monly found in microbenchmarks [3, 33, 39, 40, 42] We believe there exists a strong duality between microbenchmarks and gray box techniques. First, ICLs often require that underlying components be benchmarked to configure internal thresholds and parameters. Second understanding the behavior of ICLs requires understanding the behavior of ....

[Article contains additional citation context not shown here]

D. Blackwell and M. A. Girshick. Theory of Games and Statistical Decisions. John Wiley & Sons, 1954.

....increases. Usually, uncertainty in the resource costs is modeled by assuming that the resource costs are random variables with fixed probability distributions, which may or may not be known to the network [2] From decision theoretic perspective this approach lies within Bayesian framework [3]. However, even when the forms of the probability distributions can be reliably identified, e.g. exponential distributions for the delays, the parameters of these distributions, e.g. average delays, remain to be subject to uncertainty within the corresponding confidence regions, leaving the ....

D. Blackwell and M. Girschick, Theory of Games and Statistical Decisions, Wiley, New York (1954).

....using algorithmic knowledge of the OS, a gray box ICL may be able to interact with the OS in a more e#cient manner. Determining how to interact with a component given only general knowledge of how it behaves has been studied extensively in theoretical work such as game theory and decision theory [9, 41]. On the practical side, there exists a tension between the optimizations one makes in an ICL and its portability: the more algorithmic knowledge that is assumed, the more optimizations one can make, but the fewer systems to which those assumptions may apply. Monitor outputs. Given only ....

D. Blackwell and M. A. Girshick. Theory of Games and Statistical Decisions. John Wiley & Sons, Inc., New York, 1954.

....or follow some defined specification. Much theoretical work has investigated the problem of making decisions with only general knowledge of how other components behave; at a high level, game theory [30] predicts the actions of adversaries in the absence of any communication, while decision theory [9] optimizes the utility of decisions based on uncertain information that has been quantified with a probability measure. As a specific example, in the LARD load balancing web server presented in [31] the front end leverages its algorithmic knowledge of how back end servers cache web pages to ....

D. Blackwell and M. A. Girshick. Theory of Games and Statistical Decisions. John Wiley & Sons, Inc., New York, 1954.

.... (i.e. the smallest of real numbers ff, for which a ffi Gammainput uncertainty leads to a ff Gammaoutput error) In order to compare different operations in terms of robustness, we will proceed as in standard decision theory, where r f (ffi) plays the role of the risk (see, e.g. [2]) Definition 4. We say that operations f(a; b) and g(a; b) are equally robust if for every ffi, r f (ffi) r g (ffi) We say that an operation f(a; b) is more robust than an operation g(a; b) if for every ffi, r f (ffi) r g (ffi) and at least for one ffi 0, r f (ffi) r g (ffi) ....

Blackwell, D. and Girshick, M.(1979). Theory of Games and Statistical Decisions. Dover, N. Y.

.... (i.e. the smallest of real numbers ff, for which a ffi Gammainput uncertainty leads to a ff Gammaoutput error) In order to compare different operations in terms of robustness, we will proceed as in standard decision theory, where r f (ffi) plays the role of the risk (see, e.g. [BG79]) Definition 4. We say that operations f(a; b) and g(a; b) are equally robust if for every ffi, r f (ffi) r g (ffi) We say that an operation f(a; b) is more robust than an operation g(a; b) if for every ffi, r f (ffi) r g (ffi) and at least for one ffi 0, r f (ffi) r g (ffi) ....

D. Blackwell and M. A. Girshick. Theory of games and statistical decisions. Dover, N. Y., 1979.

....states that if the risk set is bounded, then inf d2D sup 2 Gamma r( d) sup 2 Gamma inf d2D r( d) 5 and there exists a least favorable distribution 0 . If in addition the risk set is closed, then there exists a minmax decision rule d 0 , and it is Bayes with respect to 0 . See Blackwell and Girshick (1954, Theorem 2.4.2) and Ferguson (1967, Theorem 1, p. 82) We shall assume that the risk set is closed and bounded. The first step in our algorithm is to find a Bayes rule with respect to a given prior distribution . Note that r( d) Z Z Z Theta L( z; d(z) f(z j ) d( d(z) Z Z [ ....

....to satisfies d (z) arg min a2A Z Theta L( z; a) d( j z) 3) where is the posterior distribution of conditional on Z: B j z) Z B f(z j ) d( OE Z Theta f(z j ) d( 4) The optimal choice under minimizes the posterior expected loss. See Wald (1950, chap. 5. 1) Blackwell and Girshick (1954, chap. 7.3) and Ferguson (1967, chap. 1.8) If the minimizer in (3) is not unique, then a Bayes rule could randomize over the set of minimizers. Let M J denote the J Gamma 1 dimensional simplex: M J = fffi 2 R J Gamma1 : ffi j 0; J Gamma1 X j=1 ffi j 1g; and let ffi denote the ....

Blackwell, D. and M. Girshick (1954), Theory of Games and Statistical Decisions, New York: Wiley.

....Meeting in September 1948 under the old title Statistics and the Theory of Games . David Blackwell (b. 1919 at Centralia, Illinois) mathematician at Howard University, and the statistician at Stanford M. Girshick later wrote together the book Theory of Games and Statistical Decisions (see [15]) But the new title and the following sentence is the personal style of Kenneth J. Arrow (b. 1921 in New York City) It may be remarked that the problem of optimum sequential choice among several actions is closely allied to the economic problem of the rational behavior of an entrepreneur under ....

Blackwell, D. and Girshick, M.A., 1954. Theory of Games and Statistical Decisions. Wiley, New York.

....model the knowledge of the opponent offering a bet to an agent at a given point in the run. One obvious choice is to assume you are playing against someone whose knowledge is identical to your own. This is what decision theorists implicitly do when talking about an agent s posterior probabilities [BG54]; it is also how we can understand the choice of probability space made in [FZ88] By way of contrast, the choice in [HMT88] corresponds to playing someone who has complete knowledge about the past and knows the outcome of the coin flip; this corresponds to the viewpoint that says that when the ....

D. Blackwell and M. A. Girshick. Theory of Games and Statistical Decisions. John Wiley & Sons, Inc., New York, 1954.

....multiple decision makers and one or more loss functionals (or performance criteria) In this case, its use is much more general than a game in the intuitive sense. The formulations 18 presented in this dissertation share similarities with concepts from statistical decision theory (e.g. [20], 46] 45] optimal control theory (e.g. 28] 105] 202] dynamic noncooperative game theory [6] and games considered in artificial intelligence research (e.g. 149] 198] The amount of cooperation that occurs between decision makers in a game is one of the key differences between ....

....yielding a pure strategy or deterministic strategy, or it could be specified with probability distributions over the sets of actions, yielding a mixed strategy or randomized strategy. The effects of noise or disturbances can be modeled by the addition of a special decision maker called nature [20]. One of the primary interests in this dissertation is 20 the modeling and incorporation of uncertainty into a robot strategy. Nature might be assumed to have a known, randomized strategy that interferes with the efforts of the decision makers. Nature could alternatively be assumed to implement a ....

[Article contains additional citation context not shown here]

D. Blackwell and M. A. Girshik. Theory of Games and Statistical Decisions. Dover Publications, New York, NY, 1979.

....are satis. ed. 2 We end this section with some comments on related mathematical results. There are many results similar to the equivalence of (i) and (iii) with continuity instead of the fair price condition and with an invariance condition for scalar multiplication (homotheticity) added ([5] Theorem 4.3.1, 28] 33] 35] 48] Additivity of preference amounts to commutativity of an ordering and an addition operation which is extensively studied in the mathematics literature ( 4] Chapter 15, 18] 24] Section 6 2.2.5) These results often consider more general state spaces and ....

Blackwell, David & M.A. Girshick (1954), "Theory of Games and Statistical Decisions. " Wiley, New York.

.... that each player uses in deciding his or her action in the market from one week to the next, using a framework outlined 352 Robert Marks 3 The concept of partitioning in order to use the coarsest (or minimal) partition which is as informative as the non partitioned space was introduced by Blackwell Girshick (1954). by Lipman (1995) Each week, faced with the actual external state (or E state) the player perceives an internal state (or P state) which will update his beliefs, on which are conditioned his action for the week, which together with the actions of his strategic rivals determines his profit ....

Blackwell, D., and Girshick, M. A., 1954. Theory of Games and Statistical Decisions, John Wiley, New York.

....proof, we have seen that for a potential, fl locally extremal at p implies fl locally constant around p. This iextremum principlej holds due to the de nition of our potentials as weighted averages of (Doob, 1984) Functions similar to our potentials also appear in statistical decision theory (Blackwell Girshick, 1954) (Borovkov, 1987, ch.V) and in game theory (Von Neumann Morgenstern, 1953) In the last case they are used to represent numerically the economic notion of utility. Remark 2.6 Except for very special , the potential of will be non convex on Gamma(p) as soon as dim(p) 2. For instance, in ....

Blackwell, D., & Girshick, M. 1954. Theory of games and statistical decisions. Wiley.

....system. For example, game theory focuses on predicting the actions of adversaries in the absence of any communication [129] Similarly, decision theory provides a model for optimizing the utility of decisions based on uncertain information that has been quantified with a probability measure [15]. Finally, team decision theory combines the previous two approaches, stressing the distributed nature of the decision makers [114] Components can also infer the state of remote components if they have common knowledge [71] Common knowledge is the strongest form of knowledge in a distributed ....

D. Blackwell and M. A. Girshick. Theory of Games and Statistical Decisions. John Wiley & Sons, Inc., New York, 1954.

....double indices to indicate an infinite sequence of copies of each type m = 1; M , we shall continue to write X = fX 1 ; XM g for a generic set of tests. This is precisely the set up in [14] More generally, this setting is at the intersection of sequential statistics [11] game theory [5] and adaptive control processes [4] In these domains, optimal strategies can, in principle, be computed using dynamic programming; still, cases in which they can be expressed in simple analytic terms are uncommon and the emphasis is on asymptotic results (e.g. Ed(T ) 1) for greedy procedures. ....

D. Blackwell and M. A. Girschick. Theory of Games and Statistical Decisions. John Wiley, 1954.

No context found.

D. Blackwell and M. A. Girshnik. Theory of Games and Statistical Decisions. Dover Publications, New York, NY, 1979.

No context found.

D. Blackwell and M. A. Girshik. Theory of Games and Statistical Decisions. Dover Publications, New York, NY, 1979.

No context found.

D. Blackwell and M. A. Girshick [1954], Theory of Games and Statistical Decisions, Wiley, New York, NY.

No context found.

D. Blackwell and M. A. Girshick. Theory of Games and Statistical Decisions. John Wiley & Sons, Inc., New York, 1954.

No context found.

D. Blackwell and M. A. Girshik. Theory of Games and Statistical Decisions. Dover Publications, New York, NY, 1979.

No context found.

D. Blackwell and M. Girschick, Theory of games and statistical decisions, Wiley, New York, NY, 1954.

No context found.

Blackwell D. & Girshick M.A. 1954, Theory of Games and Statistical Decisions, New York: John Wiley.

No context found.

D. Blackwell and M. A. Girschick, Theory of Games and Statistical Decisions (Wiley, New York, 1954).

No context found.

Blackwell D. & Girshick M.A. 1954, Theory of Games and Statistical Decisions, New York: John Wiley.

No context found.

D. Blackwell and M. A. Girshick. Theory of games and statistical decisions. John Wiley, New York, 1954.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC