Rich, E.; Knight, K. Artificial Intelligence. McGraw Hill, 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and the logic approach is elegant and has a pure semantic. The aim of this paper is to present a new logic architecture for intelligent agents (LASG a Logic Architecture based on Stacks of Goals) This architecture combines the traditional logic architecture with a planning architecture [3]. The advantages of the proposed architecture are shown in the paper. Keywords: intelligent agents, logic. 1. Introduction The logic approach is a topic of Symbolic Artificial Intelligence and has its own importance in the field of intelligent agents, even if it is well known the controversy ....

....(Xf, Yf) the non deterministic predicate solution(Xi, Yi, L1, L, M, N) i, i, i, o, i) collects the elements of a solution in L (L1 is the former generated list) The goal has the form goal: path(4, 1, 1, 3, 4, 3, L) and the solutions are two: Figure 4. The Prolog program L= 4, 1] [3, 1], 2, 1] 2, 2] 1, 2] 1, 3] 6. Conclusions and further work In a logic based architecture, the intelligent behavior is generated by a symbolic representation of the environment and the agent s behavior, and by a symbolic manipulation of this representation. In the logic based approach, ....

[Article contains additional citation context not shown here]

Rich, E., Knight, K. "Artificial Intelligence", Mc Graw Hill, New York, 1991

....or assists in the presence of an manager. The literature provided in the special issues of the journals [1] 2] 3] 4] the collection of papers provided in [5] 6] and the text books [8] 9] would be of great help for those interested in knowing the current trends in NM. Artificial Intelligence (AI) [10] is a branch of Computer Science (CS) that includes such automated programs in the form of an Expert System (ES) which are expected to mimic an expert of one particular domain. An idea to deploy such expert systems to assist the network managers motivates the use of artificial intelligence methods ....

E. Rich, Artificial Intelligence, McGraw-Hill, Singapore, 1983.

.... choice with the same consequences for its fourth object and therefore finish with a suboptimal cache of value 9 (Figure 3b) or 8 (Figures 3c and 3d) Simulated Annealing (SA) Simulated Annealing (SA) is a randomized algorithm that statistically simulates the slow cooling of a physical system [22]. The algorithm works by choosing an initial state (the current state ) within a state space consisting of all possible solutions, and then performing a random walk beginning at this state. The walk consists of choosing a random neighbor, and proceeding to make this neighbor the current state if ....

E. Rich and K. Knight. Artificial Intelligence. McGraw-Hill, 1991.

....plan which can be followed based on the sensor data collected during the orientation process. The time complexity terms (T A (k, n) TB (k, n) etc. are from [AM99] except when noted. 5. 1 Algorithm for Optimal Length Plans For a fine description of AND OR search and the similar AO algorithm, see [RK91]. Each level of the search tree corresponds to a push align operation. The root node of the tree is the set of all possible orientations of the part. A node in the tree corresponds to the set of possible states consistent with the sequence of push align operations and sensor readings up to that ....

E. Rich and K. Knight. Artificial Intelligence. McGraw Hill, 1991.

....contains a variable. For example, an answer to the question Who went to the party might be a list of individuals who meet the criterion went to the party or a description [Green and Raphael, 1968, Green, 1969b, Green, 1969a, Luckham and Nilsson, 1971, Reiter, 1978a, Nilsson, 1980, McCune, 1994, Rich and Knight, 1991, Russell and Norvig, 1995, Luger and Stubblefield, 1997] A ground term, or fact, is a variable free term. of these individuals. The fact that David went to the party may be used to complete a proof that someone went to the party, but the answer provided by a traditional theorem prover under ....

....or lists of facts. With the emergence of theorem proving as a dominant paradigm for question answering, specific answers became the focus of question answering systems. The answer is a list of facts dogma has proven remarkably persistent in mainstream AI. Presentations in standard textbooks on AI [Rich and Knight, 1991, Ginsberg, 1993, Russell and Norvig, 1995, Luger and Stubblefield, 1997] perpetuate this view. Allen s book on natural language understanding refers exclusively to the retrieval of facts in its discussion of question understanding [Allen, 1995] In Russell and Norvig s (1995) text on artificial ....

[Article contains additional citation context not shown here]

Elaine Rich and Kevin Knight. Artificial Intelligence. McGraw Hill, 2 edition, 1991.

.... tracking and plan recognition are both part of a larger family of comprehension capabilities that enable an agent to parse a continuous stream of input from its environment, whether it be in the form of natural language or speech or music or simulated radar input, as is the case here (e.g. see [ Rich and Knight, 1990, chapter 14 ] Resolving ambiguities in the input stream is a key problem when parsing all of these different types of input. One example of the ambiguity faced in agent tracking can be seen in L s turn in Figure 1 c. From D s perspective, L could be turning to fire a missile. Alternatively, L ....

E. Rich and K. Knight. Artificial Intelligence. McGraw-Hill, New York, NY, 1990.

....development of semantic networks starting from a predicate logic description. This aspect became more important with the success of knowledge based systems. And indeed the eighties, the high time of knowledge representation, led to AI books focusing on various knowledge representation formats. In [Rich1983] and [Charniak and McDermott1985] considerable parts deal with knowledge representation formats like frames, conceptual dependencies or semantic networks. There was a clear shift from general problem solving techniques, like heuristic search and inference, towards knowledge representation and its ....

Elaine Rich. Artificial Intelligence. McGraw-Hill, 1983.

....0 ) where, for any x E and X E, K 2 = 1, k 2 with k 2 = max d(x, y ) x, y E and d(x, X) min #d(x, y) 1 2# : y X , where d(x, y) is the Euclidean distance between x and y and ### is the greater integer number lesser than #. 2.3. Expert Systems. Expert systems [28] use the knowledge of an expert in a given specific domain to answer non trivial questions about that domain. For example, an expert system for image classification would use knowledge about the characteristics of the classes present in a given region to classify a pixel in an image of that ....

Rich, E.; Knight, K. Artificial Intelligence. McGraw Hill, 1993.

.... with human thinking, activities such as decision making, problem solving, learning [3] Thinking Rationally: The study of mental faculties through the use of computational models [5] Acting Humanly: The study of how to make computers do things at which, at the moment, people are better [55]. Acting Rationally: The branch of computer science that is concerned with the automation of intelligent behavior [45] Although I have left the taxonomy as it was, I propose to switch the contents on the first two items, based on all I have said on rationality. If we regard rationality as an ....

Rich, E., Knight, K., Artificial Intelligence, 2 n edition, McGraw-Hill, 1991.

.... problem solvers (Langley, 1983, Mitchell et al. 1983) or other general purpose linear problem solvers (Etzioni, 1993, Fikes et al. 1972, Leckie and Zukerman, 1991, Minton, 1988, Prez and Etzioni, 1992) These problem solvers are known to be incomplete and incapable of finding optimal solutions (Rich, 1983, Veloso, 1989) If we remove the linearity assumption, we are dealing with nonlinear problem solvers. In this article we show that nonlinear problem solving offers new learning opportunities where domain dependent control knowledge can be used to further improve not only the problem solver s ....

Elaine Rich. Artificial Intelligence. McGraw-Hill, Inc., 1983.

.... graph without the need for serialising it [10] Further, there has been recently a renewed interest in search based planning techniques, as these have demonstrated significant performance on various planning tasks [10] 11] 12] 13] We thus use a real time variant of the AO algorithm [9] 14] [15] to search the AND OR graph. The AO algorithm is a heuristic We could refer to this interaction as the cross product of the individual characters plans. search algorithm operating on AND OR graphs: it can find an optimal solution sub graph, which in our case corresponds to a given character s ....

Knight, K. and Rich, E., 1991. Artificial Intelligence, 2nd Edition. McGraw Hill.

....from the players (see figure 1. 3) Typical examples of extensive games are different kinds of board games such as Chess, Go, and Othello and there are several algorithms dealing with the search of the states with the highest utilities, e.g. the well known minimax algorithm (with variations) [75, 79]. Another example is the centipede game, in which each player, when she is in turn, can choose between to Rock Paper Scissors Rock 0,0 1,1 1, 1 Paper 1, 1 0,0 1,1 Scissors 1,1 1, 1 0,0 Table 1.1: The payoffs for the players in Rock, Paper, Scissors. The entries show the payoffs for (player 1 ....

E. Rich and K. Knight. Artificial Intelligence. McGraw-Hill, 1991.

....produce a solution. Further, there has been recently a renewed interest in search based planning techniques, as these have demonstrated significant performance on various planning tasks [2] 10] 17] 23] As the task network for the characters is an AND OR graph, we naturally use the AO algorithm [9][15] 16] to produce a solution. The solution takes the form of a sub graph (rather than a path like in traditional graph search) In our context, the terminal nodes of this sub graph correspond to a sequence of actions that constitute a specific instantiation of the storyline. These terminal ....

Knight, K. and Rich, E., 1991. Artificial Intelligence, 2nd Edition. McGraw Hill.

....about the internal state of other agents and their goals. In order to satisfy a goal, an agent may play different roles, authorize or delegate some tasks to other agents. Hence, an agent must be able to decompose a given goal into subgoals and to assign this subgoals to the various other agents [RK91] 3 Object Specification We base our presentation on the concept of object as introduced in [SSE87] An object is an observable process. The dynamic behaviour of objects is described by life cycles built from basic events which may modify the object state. These events can be seen as an ....

E. Rich, K. Knight. Artificial Intelligence. McGraw-Hill, 1991.

....Figure 2. A Sigmoid Based Artificial Neuron The sigmoid activation function can be adapted whenever binary output is desired. As is the case in electronic gates, a certain range of output values can represent a logic 1 and another range a logic 0. The following scheme is adopted in this paper [3]: Eq. 7) The above is only enforced when binary output is desired. This usually occurs at the output layer of neurons. It is to be noted that the actual output of a neuron is not replaced by the equivalent binary value. The above equation is only used to get the equivalent binary output of a ....

....when binary output is desired. This usually occurs at the output layer of neurons. It is to be noted that the actual output of a neuron is not replaced by the equivalent binary value. The above equation is only used to get the equivalent binary output of a neuron for classification purposes [3]. Multilayer Neural Network An Artificial Neural Network (ANN) is constructed of artificial neurons. A multilayer neural network consists of an input layer, at least one hidden layer, and an output layer. Figure 3 shows a three layer neural network in which there is only one hidden layer. The ....

[Article contains additional citation context not shown here]

) Rich, E., and Knight, K., "Artificial Intelligence," McGraw-Hill, 1991.

....such as inclusive CPU utilization and wall clock time, limiting its attractiveness as the only source of performance data. Deep Start leverages the advantages of both sampling and search in the same automated performance diagnosis tool. Most introductory artificial intelligence texts (e.g. [16]) describe heuristics for reducing the time required for a search through a problem state space. One heuristic involves starting the search as close as possible to a goal state. We adapted this idea for Deep Start, using stack sample data to select deep starters that are close to the goal states ....

Rich, E., Knight, K.: Artificial Intelligence. McGraw-Hill, New York (1991)

....AI had severed practically all ties with pattern recognition, which was very counter productive to the development of both areas and particularly to AI. 1 With the recent rise of connectionism, this situation has begun to change, which is reflected in the content of the recent AI textbooks ([39], 49] As a result, another issue was brought to the fore: the existing learning models were found grossly inadequate. The present collection of papers is, basically, supposed to address the latter situation. In this chapter we shall briefly outline the answer that was originally proposed in ....

RICH, E., and KNIGHT, K. (1991). Artificial Intelligence, second edition, McGraw-Hill, New York.

....elements of top down and bottom up control: bottom up when a new frame is activated or from the tension of a mismatched frame, top down when a frame activates its terminal demons. Similar ideas put forward around the same time are Carl Hewitt s actors [18] 19] 40] and Roger Schank s scripts [29][31] Draper et al.[14] recently developed SCHEMA using a frame based structure for low level vision applications. In this system there are concur3 rent processes communicating through a blackboard, with structure matching and forward chaining. This system is used to segment natural 2D images, but ....

Elaine Rich and Kevin Knight. Artificial Intelligence. McGraw-Hill, New York, 2nd edition, 1991.

....solution would require either too many resources, or, the implementing process should be completed with numerous shortcuts. The narrowed focus, and thus, very limited application domain result in difficult compromises when considering handling of free form textual input from individual persons [Win92, Ric91, Rau96]. One additional solution to the classification problem would be the introduction of www forms, which could have enough machine computable information in them so that the processing would be relatively easy and inexpensive. This, however, is not very flexible solution because forms do not ....

Elaine Rich, Kevin Knight, "Artificial Intelligence", McGraw-Hill, 1991

....solution would require either too many resources, or, the implementing process should be completed with numerous shortcuts. The narrowed focus, and thus, very limited application domain result in difficult compromises when considering handling of free form textual input from individual persons [Win92, Ric91, Rau96]. One additional solution to the classification problem would be the introduction of www forms which could have enough machine computable information in them so that the processing would be relatively easy and inexpensive. This, however, is not a very flexible solution because forms do not ....

Elaine Rich, Kevin Knight, "Artificial Intelligence", McGraw-Hill, 1991

....[5] Each of these projects use program counter sampling as its primary technique for obtaining information about the application under study. In contrast, the Deep Start search strategy collects samples of the entire execution stack. Most introductory artificial intelligence texts (e.g. [17]) describe heuristics for reducing the time required for a search through a problem state space. One heuristic involves starting the search as close as possible to a goal state. We adapted this idea for the Deep Start search strategy, using stack sample data to select deep starters that are close ....

E. Rich and K. Knight, Artificial Intelligence, McGraw-Hill, Inc., 1991.

....other than the variable sharing property defined above. Second, the definitions provided for generic and hypothetical answers completely clarify the way in which clauses corresponding to rules should be classified: rule and generic answer are not the same thing. 3 Examples A simple example from Rich and Knight [1991, page 192] is used as a first illustration of hypothetical answering. Consider the following knowledge base (KB) 1. Cats like to eat fish 2. Cats like to eat tuna 3. Calicos are cats 4. Herb is a tuna and the question What does Boots like to eat What sorts of answers arise according to the ....

Elaine Rich and Kevin Knight. Artificial Intelligence. McGraw Hill, 2 edition, 1991.

....composed of subtasks that execute in parallel, such that the desired behavior emerges [16] For example, the driving task in task specification 3 2. The reader should note that the discussion of serial and parallel execution of tasks does not mean that the task decomposition is an and or tree [64]. Task decompositions used in this methodology do not have or branches, though a particular subtask may select among various choices when deciding what to do. Task: drive car to the store Environment: car on road with other cars, lane markings, street signs, traffic lights, start in ....

Rich, E. and Knight, K. 1991. Artificial Intelligence. McGraw-Hill. New York, NY.

.... of logical programming languages like Prolog (Clocksin and Mellish 1994) in cognitive science, it is the foundation of a number of category based approaches to the analysis of natural language (Shieber 1986) The basic unification algorithm can be found in many introductory AI textbooks (e.g. Rich and Knight 1991 p. 152) and can be summarized recursively as follows: 1) A variable can be unified with a literal. 2) Two literals can be unified if their initial predicate symbols are the same and their arguments can be unified. If, for example, we have a Prolog database containing the assertion ....

Rich, E. and K. Knight (1991). Artificial Intelligence. New York: McGraw-Hill.

....Tabu search uses a list called the Tabu list which dynamically changes throughout the course of the search and contains a group of the most recently visited points of the search space. The search avoids going back to these states so as not to keep cycling. Dependency directed backtracking [ Rich and Knight 1991 ] is a search method similar to depth first search which uses search history to decide which state to backtrack to in case a search path proved fruitless. Explanation based learning (EBL) Mitchell 1997 ] takes the outcome of a training process (such as a rule, a proof or a decision tree) ....

Elaine Rich and Kevin Knight. Artificial Intelligence. McGraw-Hill, 1991.

....its corresponding BAG is shown in Figure 3 and Figure 4, respectively. The normal mode and test mode bridges are shown in Figure 3 as shaded and unshaded triangles. In this example n = 6 and m = 22. The index of bridges are written near each edge in Figure 4. Applying a depth first search (DFS) [Rich93] to the BAG can easily identify all paths to from SoC ports. Specifically, starting at an output port and following edge directions, the DFS lists all paths from core outputs to the SoC s output ports. Similarly, starting at an input port and following the edges in opposite directions, the DFS ....

....directions, the DFS lists all paths from core outputs to the SoC s output ports. Similarly, starting at an input port and following the edges in opposite directions, the DFS lists all paths from the SoC s input ports to the core input ports. We applied the classical DFS algorithm discussed in [Rich93]. However, an important modification is needed to guarantee a valid TAM solution. If in the search process, we revisit a node (port bus) or an edge P4 P3 P1 P2 P6 P5 B1 B2 IO1 IO2 1 2 4 7 9 10 3 5 11 12 14 15 19 18 16 20 21 22 8 17 13 6 Figure 4: The Bridge Access Graph ....

[Article contains additional citation context not shown here]

E. Rich, Artificial Intelligence, McGrawHill Inc., 1993.

....and pragmatics are used to perform the understanding. Indeed, Minsky s quote above seems to summarize the opinions of many AI researchers with respect to statistical modeling. Elaine Rich mentions that understanding is the process of converting an input sentence from one representation into another[72]. In particular, a user interacting with a computer expects to receive quick and accurate information in response to a query. The creation of this response, either by another person or a program, demonstrates the responder s understanding of the user s query. If a program or another human ....

....a few comments are in order. Generally, the designer of an NLU system writes a decoder that searches the meaning space for the formal language that maximizes equation 2.2. The most common search procedures are the stack search[33] and the Viterbi beam search[44] both variants of the A search[72]. As the decoder searches the space of F , the translation model parameters 37 are used to calculate either the maximum likelihood value: p Theta (E j F ) X A p Theta (A; E j F ) 2.69) or the Viterbi value: p Theta (E j F ) maxA p Theta (A; E j F ) 2.70) This translation model ....

Elaine Rich. Artificial Intelligence. McGraw-Hill, New York, NY, 1983.

....follows c = Z t min (d)j(x GammaX (d) y GammaY(d) jt max (d) 1dxdy where the integral must be calculated over all points inside the field. As will be seen, we need to change this formula slightly. Algorithm 1 illustrates outline of the method. This algorithm follows the hillclimbing approach [2]. In this algorithm, by H(x; y) we mean the set of neighbors of (x; y) As one can observe, the function ffi d as we defined for server can not be applied to this algorithm, because this function has a flat behavior over a wide area on its Input: I: Set of input data d: Positioning accuracy. ....

E. Rich, K. Knight, Artificial Intelligence.

.... attr, city ) Delta Delta Delta g 3.3 Query Frame In the natural language query processing, a structure is necessary for effectively expressing and managing the knowledge. Many AI applications use the frame structure because it is a simple but powerful tool to store and process the knowledge[10]. Thus the KID adopts the frame based approach to express and manipulate the knowledge of the transformation process and the query frame is introduced to represent the knowledge. A query frame consists of the name of the frame, the parents of the frame, and six slots of the frame and their values. ....

E. Rich and K. Knight, Artificial Intelligence, McGraw-Hill, Inc., 1991. 14

No context found.

Rich, E.; Knight, K. Artificial Intelligence. McGraw Hill, 1993.

No context found.

Rich, Alaine and Knight, Kelvin, 1991. Artificial Intelligence, 2e. Tata McGraw-Hill Publishing Company Limited, New Delhi

No context found.

Rich, E., and Knight, K. 1991. Artificial Intelligence. 2d ed. New York: McGraw-Hill.

No context found.

Rich, E., & Knight, K. (1991). Artificial intelligence. McGraw-Hill.

No context found.

Rich E, Knight K. (1991) Artificial Intelligence. 2nd ed, McGraw-HillInc, New York.

No context found.

Rich, E. ; Knight, K. (1983), Artificial Intelligence, McGraw-Hill.

No context found.

E. Rich and K. Knight, Artificial Intelligence. McGraw-Hill, 1992.

No context found.

E. Rich and K. Knight, Artificial Intelligence. McGraw-Hill, 1992.

No context found.

Elaine Rich and Kevin Knight. Artificial Intelligence. McGraw-Hill, 2nd edition, 1991.

No context found.

Rich, E., and Knight, K., Artificial Intelligence, Second ed., McGraw-Hill, 1992.

No context found.

Elaine Rich and Kevin Knight. Artificial Intelligence. McGraw-Hill, 1990.

No context found.

Elaine Rich and Kevin Knight. Artificial Intelligence. McGraw-Hill, Inc., second edition, 1991.

No context found.

E. Rich, K. Knight, Artificial Intelligence, McGrawHill, 1991.

No context found.

E. Rich and K. Knight. Artificial Intelligence. McGraw-Hill, Inc., 2 edition, 1991.

No context found.

E. Rich and K. Knight. Artificial Intelligence, 2nd Ed. McGraw-Hill, 1991.

No context found.

E. Rich. Artificial Intelligence. Singapore, McGrawHill, Inc., 1986.

No context found.

Rich, Alaine and Knight, Kelvin, 1991. Arti- ficial Intelligence, 2e. Tata McGraw-Hill Publishing Company Limited, New Delhi

No context found.

Elaine Rich and Kevin Knight. Artificial Intelligence. McGraw-Hill, New York, second edition, 1991.

No context found.

E. Rich and K. Knight. Artificial Intelligence. McGraw-Hill, New York, 2nd edition (1991). 37

No context found.

Rich, E., and Knight, K. Artificial Intelligence, 2nd ed. McGrawHill, New York, 1991.

No context found.

Rich, E. and Knight, K. (1981) Artificial Intelligence. McGraw-Hill, New York.

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