Training and holistic computation of vector graphics with Hebbian bases in contrast to RAAM networks

Hebbian Learning is well-known for training of associative networks whereas recursive auto-associative memory (RAAM) learning uses auto-associative networks which are trained to represent structured information like parse trees of natural sentences or logical terms. In this paper Hebbian learning is used for representing structured information in terms of vector graphic. The resulting networks are holistically computed. Furthermore, a theorem relating bipolar Hebbian learning is proved.