|
"
Towards active and progressive learning in Artificial Neural Networks"
Download: |
This paper proposes a broad overview of the use of a learning strategy in the field of supervised neural networks.
By learning strategy we mean a technique that handles the order of presentation of the patterns to be learned.
We distinguish two types of works: informative learning and progressive learning, which are presented in the two first sections. In the third section, we introduce the concept of active and progressive learning as a challenge for further research. New algorithms based on this principle, that are still to be designed, should significantly improve the learning abilities of present learning algorithms |
(( TOPICS
Progressive learning (( LABS Lifia |