Learning Goals in Multistrategy Learning- GMU Machine Learning and Inference Laboratory

learning goals in multistrategy learning

(Michalski, Utz)

Learning arises from an intelligent individual's inability to reason and comprehend with its current knowledge. From prior research, every learning task requires background knowledge, sufficient inputs and a learning goal to achieve success. In multistrategy learning, though, the pupil faces additional complexity: here, the pupil must derive several learning goals for sequential and possibly parallel application in the learning process. The pupil must use these goals then to choose relevant inputs and necessary strategies (among several) in a timely way in order to acquire the "right" target knowledge.

The aims of this research project are twofold: (1) to experiment with a formalism to specify goals for multistrategy learning and (2) to construct a processing mechanism to generate and apply learning goals appropriately. This project is based on the MTL methodology supporting the Inferential Theory of Learning. Such a formalism must be accurate and complete to enable the processing mechanism to create explicit learning goals to understand the context, direct the procedure, and evaluate the results of multistrategy learning tasks. The formalism must be domain-independent as well as task-adaptive.

Proper specification is essential for success. Learning goals must be specified to enable an examination of any newly acquired or more efficient knowledge. When the examination indicates that the results are implausible or incompatible with the target knowledge, the process must be capable of trying again. It should reselect inputs or learning strategies on the advice of the original learning goals or, where necessary, regenerate alternative learning goals to redirect the learning task.

Selected References

Michalski, R.S. and Ram A., "Learning as Goal-Driven Inference," in Goal-Driven Learning, A. Ram & D. B. Leake (Eds.), MIT Press/Bradford Books, Cambridge, MA, 1995.

For more references, seeĀ publications section.

MLI Copyright © 2017 Machine Learning and Inference Laboratory
College of Health and Human Services, George Mason University
4400 University Dr, MSN 1J3, Fairfax, VA 22030, U.S.A