Milind Tambe, Robert Doorenbos, and Allen Newell. The match cost of adding a new rule: a clash of views. Technical Report CMU-CS-92-158, Carnegie Mellon University, Pittsburgh, PA 15213, 1992. 43 Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Learning Analytically and Inductively - Mitchell, Thrun (1995) (2 citations) (Correct)
....a form of analytical learning, and the screening effect of new learned rules on inconsistent prior knowledge allows for an inductive component to Soar s learning. Whereas the capabilities of Soar s analytical learning mechanism have been convincingly demonstrated in a variety of domains (e.g. [Tambe et al. 1992]) its inductive learning capabilities have yet to be demonstrated on the same scale. Our assessment is that while combining inductive and analytical learning will be important to the goal of Soar as a general model of learning, we foresee two difficulties with the current strategy of combining ....
....where a complete rule is formu lated based solely on the analysis of a single example. Others have noted [Rosenbloom, 1983] that learning in Chunking Soar may be too abrupt to be a correct model of human learning. EBNN Soar may avoid the average growth effect encountered by Chunking Soar [Tambe et al. 1992], in which the large number of learned chunks can lead to signifi 35 cant slowdown in processing. By representing knowledge using a bounded collection of neural nets (one per impasse, problem space pair) the cost of applying such knowledge cannot grow indefinitely. Because it would use a ....
Milind Tambe, Robert Doorenbos, and Allen Newell. The match cost of adding a new rule: a clash of views. Technical Report CMU-CS-92-158, Carnegie Mellon University, Pittsburgh, PA 15213, 1992. 43
Learning Analytically and Inductively - Mitchell, Thrun (1995) (2 citations) (Correct)
....a form of analytical learning, and the screening effect of new learned rules on inconsistent prior knowledge allows for an inductive component to Soar s learning. Whereas the capabilities of Soar s analytical learning mechanism have been convincingly demonstrated in a variety of domains (e.g. [Tambe et al. 1992]) its inductive learning capabilities have yet to be demonstrated on the same scale. Our assessment is that while combining inductive and analytical learning will be important to the goal of Soar as a general model of learning, we foresee two difficulties with the current strategy of combining ....
....where a complete rule is formu lated based solely on the analysis of a single example. Others have noted [Rosenbloom, 1983] that learning in Chunking Soar may be too abrupt to be a correct model of human learning. EBNN Soar may avoid the average growth effect encountered by Chunking Soar [Tambe et al. 1992], in which the large number of learned chunks can lead to signifi cant slowdown in processing. By representing knowledge using a bounded collection of neural nets (one per impasse, problem space pair) the cost of applying such knowledge cannot grow indefinitely. Because it would use a bounded ....
Milind Tambe, Robert Doorenbos, and Allen Newell. The match cost of adding a new rule: a clash of views. Technical Report CMU-CS-92-158, Carnegie Mellon University, Pittsburgh, PA 15213, 1992.
Learning Analytically and Inductively - Mitchell, Thrun (1995) (2 citations) (Correct)
....a form of analytical learning, and the screening effect of new learned rules on inconsistent prior knowledge allows for an inductive component to Soar s learning. Whereas the capabilities of Soar s analytical learning mechanism have been convincingly demonstrated in a variety of domains (e.g. [Tambe et al. 1992]) its inductive learning capabilities have yet to be demonstrated on the same scale. Our assessment is that while combining inductive and analytical learning will be important to the goal of Soar as a general model of learning, we foresee two difficulties with the current strategy of combining ....
....where a complete rule is formu lated based solely on the analysis of a single example. Others have noted [Rosenbloom, 1983] that learning in Chunking Soar may be too abrupt to be a correct model of human learning. EBNN Soar may avoid the average growth effect encountered by Chunking Soar [Tambe et al. 1992], in which the large number of learned chunks can lead to significant slowdown in processing. By representing knowledge using a bounded collection of neural nets (one per impasse, problem space pair) the cost of applying such knowledge cannot grow indefinitely. Because it would use a bounded ....
Milind Tambe, Robert Doorenbos, and Allen Newell. The match cost of adding a new rule: a clash of views. Technical Report CMU-CS-92-158, Carnegie Mellon University, Pittsburgh, PA 15213, 1992.
Learning Analytically and Inductively - Mitchell, Thrun (1995) (2 citations) (Correct)
....a form of analytical learning, and the screening effect of new learned rules on inconsistent prior knowledge allows for an inductive component to Soar s learning. Whereas the capabilities of Soar s analytical learning mechanism have been convincingly demonstrated in a variety of domains (e.g. [Tambe et al. 1992]) its inductive learning capabilities have yet to be demonstrated on the same scale. Our assessment is that while combining inductive and analytical learning will be important to the goal of Soar as a general model of learning, we foresee two difficulties with the current strategy of combining an ....
....where a complete rule is formulated based solely on the analysis of a single example. Others have noted [Rosenbloom, 1983] that learning in Chunking Soar may be too abrupt to be a correct model of human learning. ffl EBNN Soar may avoid the average growth effect encountered by Chunking Soar [Tambe et al. 1992] , in which the large number of learned chunks can lead to significant slowdown in processing. By representing knowledge using a bounded collection of neural nets (one per impasse, problem space pair) the cost of applying such knowledge cannot grow indefinitely. ffl Because it would use a ....
Milind Tambe, Robert Doorenbos, and Allen Newell. The match cost of adding a new rule : a clash of views. Technical Report CMU-CS-92-158, Carnegie Mellon University, Pittsburgh, PA 15213, 1992.
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