Français
2. Methods for coordinating learning algorithms
Overall, three main families of combinations of multiple learning algorithms have been proposed: merging the outputs of these algorithms before making a decision; selecting one of these algorithms, which then takes sole control of the agent; this selection may result from monitoring the evolution of internal variables, or from a second learning layer.
2.1 Static fusion
If we're not looking to optimize the use of computing resources, but only to improve an agent's behavior, we can systematically calculate the outputs of all the learning systems, and then merge them before making a decision. The idea is then that actions that meet with consensus are probably the best.
The first method of coordinating learning algorithms
The Ultimate Scientific and Technical Reference
This article is included in
Robotics
This offer includes:
Knowledge Base
Updated and enriched with articles validated by our scientific committees
Services
A set of exclusive tools to complement the resources
Practical Path
Operational and didactic, to guarantee the acquisition of transversal skills
Doc & Quiz
Interactive articles with quizzes, for constructive reading
Methods for coordinating learning algorithms
Conclusion
Bibliography
The Ultimate Scientific and Technical Reference
