Kohonen network / Adaptive resonance theory

Kohonen network / Adaptive resonance theory

1. computational method for the visualization and analysis of high-dimensional data

2. vector quantization

   • classical quantization technique
   • prototype vectors

3. Self-Organizing Map

   • Form of the array
     • Scaling of the vector components
       • Learning with a small number of available training samples
         • Quality of learning:
         • a finite set of attributes
       • subtle problem
     • A hexagonal grid of nodes

4. is a theory developed by Stephen Grossberg and Gail Carpenter

5. The Stability-Plasticity Dilemma and Rapid Learning Throughout Life

   • Equations
     • Modules
       • Modal Architectures
         • Complementary Processing Streams

6. Equations for short-term memory

   • STM: Neuronal Activation in Additive and Shunting On-Center Off-Surround Networks
     • MTM: Habituative Transmitter Gates and Depressing Synapses
       • LTM: Gated Steepest Descent Learning

7. Mismatch-activated Nonspecific Arousal Regulates Reset and Search

   • Complementary attentional and orienting systems
     • Novelty-sensitive nonspecific arousal: Novel events are arousing
       • Medium-term memory: Habituative transmitter gates in nonstationary hypothesis testing

Images and files in the map