Classifier with Hierarchical Topographical Maps as Internal Representation
This work addresses the problem of enhancing representational learning in AI by proposing a biologically relevant architecture, but it appears incremental as it builds on previous work without clear broad impact.
The authors tackled the challenge of integrating biologically-inspired hierarchical topographical maps into deep learning by developing a classifier with context-relevant maps that combine bottom-up and top-down learning, resulting in a model tested in a more challenging setting with advanced hidden layers.
In this study we want to connect our previously proposed context-relevant topographical maps with the deep learning community. Our architecture is a classifier with hidden layers that are hierarchical two-dimensional topographical maps. These maps differ from the conventional self-organizing maps in that their organizations are influenced by the context of the data labels in a top-down manner. In this way bottom-up and top-down learning are combined in a biologically relevant representational learning setting. Compared to our previous work, we are here specifically elaborating the model in a more challenging setting compared to our previous experiments and to advance more hidden representation layers to bring our discussions into the context of deep representational learning.