Some Experiments Around a Neural Network for Multimodal Associations

B. Yann and R. Abdelmalek (France)


Neurons networks, bidirectional associative memory


This paper presents a study of the model of triple BAM by [11] which is an improved variation of the original BAM model by [7]. This class of model aims at integrating differ ent sensory inputs in order to memorize a uniļ¬ed and dis tributed representation. An experimental evaluation of the model is presented that underlines its limitations in terms of noise robustness and learning capacities. A new model is presented in order to overcome those initial limitations by introducing a new online learning algorithm adapted from the PRLAB initial algorithm that improve both noise ro bustness and learning capacities. Finally, model proper ties and limitations are considered and discussed within the context of multi-modal integration and brain modeling.

Important Links:

Go Back