S. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of Artificial Intelligence Research, 2:575--609, 1995. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
First 50 documents Next 50
How the Semantic Web will change KR: challenges and.. - van Harmelen (2002) (Correct)
....conclusion. Furthermore, the degrees of such incompleteness or unsoundness must be a function of the available resources. Answers will often have to be approximate (where ideally, the reasoner can give us an indication of the quality of such approximations) Zilberstein and Russell, 1995, Russell et al. 1993, Lesser et al. 2000] Agent research leading KR Of course, not all of the above research challenges are new to KR, and many of them have been on the research agenda to some extent. Examples are approximate reasoning, trust, task independent formulation of knowledge, and reconciling multiple ....
Russell, S. J., Subramanian, D., and Parr, R. (1993). Provably bounded optimal agents. In Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence (IJCAI93) , pages 338--344, Chambery, France.
Real-Time Problem-Solving with Contract Algorithms - Zilberstein, Charpillet.. (1999) (9 citations) (Correct)
....control of computation by Dean and Boddy [1988] Horvitz [1988] Russell and Wefald [1991] and others. By and large, existing meta level control mechanisms are myopic; the computation is terminated once no single computational step has positive value. One exception is the technique developed by Russell, Sub ramanJan, and Parr [1993] for sequencing a set of rules given a stochastic deadline. Similar to their dynamic programming approach, our solution provides a globally optimal policy, taking into account future computations. To summarize, we show how to optimally sequence contract algorithms so as to maximize their utility ....
Stuart J. Russell, Devika Subramanian and Ron Parr. Provably bounded optimal agents. Thirteenth International Joint Conference on Artificial Intelligence, 338-344, 1993.
Optimal Sequencing of Contract Algorithms - Zilberstein, Charpillet, Chassaing (2003) (Correct)
....created by a sequence of con tracts. We want the interruptible algorithm to guarantee the same quality as the contract algorithm, if it runs on a processor that is accelerated by a factor of r 1. This definition follows the notion of bounded optimality defined by Russell, Subramanian and Parr [16]. In fact, the results presented in Section 3 can be interpreted as the construction of a bounded optimal interruptible algorithm from the contract one. Moreover, we prove that the minimal acceleration needed by any interruptible algorithm that matches the quality of J[ is 4. The acceleration ....
....to the optimization of computational utility given a set of assumptions about expected problems and constraints in reasoning resources. This definition is broader than the one used in this paper. Our notion of bounded optimality is more similar to the one used by Russell, Subramanian and Parr [16]. In particular, the latter definition focuses on optimization of program design as opposed to optimization of the actions performed by an agent. The optimal sequencing of contract algorithms is an example of program construction mechanism that fits that definition. faster processor; it is only a ....
[Article contains additional citation context not shown here]
Stuart J. Russell, Devika Subramanian and Ron Parr. Provably bounded optimal agents. Thirteenth International Joint Conference on Artificial Intelligence, 338-344, 1993.
Models of Bounded Rationality: A concept paper - Zilberstein (1995) (Correct)
....on composition of anytime algorithms. Anytime algorithms offer a tradeoff between computation time and quality of results. They have been used successfully in such applications as medical diagnosis [Horvitz and Rutledge, 1991] and real time path planning [Boddy and Dean, 1989; Zilberstein and Russell, 1993]. When anytime algorithms are combined with an appropriate monitoring scheme, they provide a powerful tool for constructing rational agents [Russell and Zilberstein, 1991; Pos, 1993] In our recent work, we have extended the benefits of gradual quality improvement over time from the level of ....
S. J. Russell, D. Subramanian and R. Parr. Provably bounded optimal agents. Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence, pp. 338-344, Chambery, France, 1993.
Monitoring And Control of anytime algorithms: a dynamic.. - Hansen, Zilberstein (2001) (13 citations) (Correct)
....We formalize the meta level control problem as a sequential decision problem that can be solved by dynamic programming, in order to construct a non myopic solution. Dynamic programming has been used before for meta level control of computation. Einav and Fehling [7] and Russell and Subramanian [27] use dynamic programming to schedule a sequence of solution methods for a real time decision making problem; and Zilberstein, Charpillet and Chassaing [40] use dynamic programming to schedule a sequence of of contract algorithms to create the best interruptible system given a stochastic deadline. ....
S. Russell, D. Subramanian, Provably bounded-optimal agents, J. Artificial Intelligence Res. 1 (1995) 1--36.
Active Logics: A Unified Formal Approach to Episodic.. - Elgot-Drapkin, Kraus, .. (Correct)
....formulas proven so far: each is believed when first proven but some may subsequently have been rejected. We will examine this definition in considerable detail later. Related work will be discussed at length in a separate section; here we simply call attention to work of Russell and Subramanian [ Russell and Subramanian, 1995 ] who discuss how one might incorporate computational limits into machine rationality concepts, and to Wellman [ Wellman, 1995 ] who notes that any formalization of a rational agent must necessarily idealize away the computational process itself. Active logic, however, is an attempt to bring ....
S. J. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of Artificial Intelligence Research, 2:575--609, 1995.
The Complexity of Agent Design Problems: Determinism and.. - Wooldridge, Dunne (2001) (Correct)
....plan representation. The obvious disadvantage is that having a positive answer to one of our agent 20 design problems (e.g. knowing that there exists an agent to carry out a task in some environment) does not imply that an agent to carry out the task will be implementable, in the sense of [13]. Most complexity results in the planning literature are bound to particular representations of goals and actions. The STRIPS notation in Bylander s work is one example [3] Baral et al. use the action description language A [2] in the work of Littman et al., the representation chosen is ST [9] ....
S. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of AI Research, 2:575--609, 1995.
The Computational Complexity of Boolean and Stochastic.. - Dunne, Wooldridge.. (2002) (Correct)
....in game theory. The obvious advantage of our approach is that our results are not bound to a particular plan representation. The obvious disadvantage is that having a positive answer to one of our agent design problems does not imply that an agent to carry out the task will be implementable (cf. [7]) Most complexity results in the planning literature are bound to particular representations of goals and actions. The strips notation in Bylander s work is one example [2] Baral et al. use the action description language A [1] in the work of Littman et al., the representation chosen is ST ....
S. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of AI Research, 2:575-609, 1995.
From Local Assessments To Global Rationality - Ekenberg, Danielson, Boman (1996) (1 citation) (Correct)
....solve different kinds of tasks and how such a system of agents can be coordinated. Not much work has been done based on results from the area of multi criteria decision aid (MCDA) cf. 3] for handling such problems. Other aspects of decision theory, however, have influenced the area of MASs (cf. [4]) partly as a result of philosophical aspects of agent rationality [5] and partly because of interest in extending the principle of maximising the expected utility in efficient real life applications [6] The viewpoint taken in this paper is that MCDA is well suited for the treatment of systems ....
S.J. Russell and D. Subramanian, Provably Bounded-Optimal Agents, Journal of Artificial Intelligence Research 2 (1995) 575--609.
The Independent Choice Logic for modelling multiple agents under.. - Poole (1997) (37 citations) (Correct)
....algorithms to build policies for agents. We want to model agents and their environments with the same language. The language should provide a decision theoretic, or game theoretic (for more than one agent) framework that can be used to build agents that can be shown to be optimal (as in [48]) or at least to have a specification of the expected utility of an agent. A planner in this framework is a program that generates a (possibly stochastic) transduction for an agent to execute. The output of the planner should be suitable for actually controlling an agent. It has to be more than a ....
....what information it knows may depend on the context, and it may be more economical to create alternatives as needed, rather than having to anticipate all of them as part of a strategy. We want to reason about the program the agent used to compute an action rather than just the action itself [48] . We want to build a representation upon a more natural specification of dynamic systems. We will extend the ICL to the dynamic ICL logic that is slightly more complicated, but arguably more natural. We model the dynamics of the world rather than the structure of the choices. The dynamic ICL is ....
[Article contains additional citation context not shown here]
S. J. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of Artificial Intelligence Research, 2:575--609, 1995.
Budgeted Learning, Part I: The Multi-Armed Bandit Case - Omid Madani Daniel Self-citation (Russell) (Correct)
....are required to match some requirements e.g. for some statistical test, or some specified and values in the case of PAC learners [Val84] We, however, have a firm total budget, specified before the learning begins. Budgeted learning falls under the framework of bounded rationality (e.g. [RS95]) and is an instance of active learning and cost sensitive learning (e.g. Ang92, CAL94, Tur00, GGR02] Feature costs in [Tur00, GGR02] refer to costs occuring at classification time, while we are concerned with cost during the learning phase. In typical poolbased active learning, the learner ....
S. Russell and D. Subramanian. Provably boundedoptimal agents. Journal of Artificial Intelligence Research, 2:575--609, 1995.
Budgeted Learning, Part II: The Naïve-Bayes Case - Lizotte, Madani, Greiner Self-citation (Russell) (Correct)
No context found.
S. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of Arti - cial Intelligence Research, 2:575-609, 1995.
Operational Rationality through Compilation of Anytime Algorithms - Zilberstein (1993) (62 citations) Self-citation (Russell) (Correct)
....agent, one can characterize its execution architecture. For example, a production system uses type rules, and a goal based system uses knowledge of type # and # . The combination of a number of execution architectures offers a tradeoff between execution time and decision quality [Ogasawara and Russell, 1993] . STATE RESULT UTILITY BEST ACTION B C D E F DT Figure 3.2: Forms of compiled and uncompiled knowledge How can intelligence be measured or evaluated Or to be more precise, given a particular agent, how can the degree of goal achievement be measured A simple approach is based on ....
S. J. Russell, D. Subramanian and R. Parr. Provably bounded optimal agents. To appear in Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence, Chambery, France, 1993.
Unknown - (Correct)
No context found.
S. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of Artificial Intelligence Research, 2:575--609, 1995.
Automatic Evaluation and Selection of Problem-Solving Methods.. - Fink (Correct)
No context found.
Stuart J. Russell, Devika Subramanian, and Ronald Parr. Provably bounded optimal agents. In Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence, pages 338--344, 1993.
From the proceedings of ECAI'98, the 13th European.. - Just-In-Time.. (1998) (Correct)
No context found.
Russell, S. J. & D. Subramanian. (1993). Provably bounded optimal agents, in the Proceedings of IJCAI93, the 13 th International Joint Conference on Artificial Intelligence, Chambery, France. San Mateo, CA: Morgan Kaufmann.
Complexity Results for Agent Design Problems - Dunne, Laurence, Wooldridge (2003) (Correct)
No context found.
S. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of AI Research, 2:575--609, 1995.
Rational Action - Lecture 3 - Limitations And Alternatives - Wooldridge, Parsons (Correct)
No context found.
Russell, S. and Subramanian, D. (1995). Provably bounded-optimal agents. Journal of AI Research, 2:575--609.
Using Performance Profile Trees to Improve Deliberation Control - Larson, Sandholm (2004) (Correct)
No context found.
Stuart Russell and Devika Subramanian. Provably bounded-optimal agents. Journal of Artificial Intelligence Research, 1:1--36, 1995.
Approximating Optimal Policies for Agents with Limited.. - Dmitri Dolgov And (2003) (Correct)
No context found.
S. J. Russell and D. Subramanian. Provably bounded optimal agents. JAIR, 2:575-- 609, 1995.
Constructing Optimal Policies for Agents with Constrained.. - Dolgov, Durfee (2003) (4 citations) (Correct)
No context found.
S. J. Russell and D. Subramanian. Provably bounded optimal agents. JAIR, 2:575--609, 1995. 20
On the Computational Complexity of Designing Bounded Agents - Laurence, Dunne, Wooldridge (2003) (Correct)
No context found.
S. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of AI Research, 2:575--609, 1995.
Near Rationality and Competitive Equilibria in Networked.. - Christin, Grossklags.. (2004) (Correct)
No context found.
S. Russell and D. Subramanian. Provably bounded-optimal agents. Journal of Artificial Intelligence Research, 2:575--609, May 1995.
Definition and Complexity of Some Basic Metareasoning Problems - Conitzer, Sandholm (2003) (Correct)
No context found.
S Russell and D Subramanian. Provably boundedoptimal agents. Journal of Artificial Intelligence Research, 1:1--36, 1995.
Definition and Complexity of Some Basic Metareasoning Problems - Conitzer, Sandholm (2002) (Correct)
No context found.
Stuart Russell and Devika Subramanian. Provably bounded-optimal agents. Journal of Artificial Intelligence Research, 1:1--36, 1995.
First 50 documents Next 50
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