A self-organizing neural network for sequence classification called SARDNET is described and analyzed experimentally. SARDNET extends the Kohonen Feature Map architecture with act...
The solving of least square systems is a useful operation in neurocomputational modeling of learning, pattern matching, and pattern recognition. In these last two cases, the soluti...
- We present an application of a massively parallel processor array VLSI circuit to the implementation of neural networks in complex architectural arrangements. The work was motiva...
David R. W. Barr, Piotr Dudek, Jonathan M. Chamber...
Models of associative memory usually have full connectivity or if diluted, random symmetric connectivity. In contrast, biological neural systems have predominantly local, non-symm...
We present an architecture for the online learning of object representations based on a visual cortex hierarchy developed earlier. We use the output of a topographical feature hier...