Premium
An Electronic Version of Pavlov's Dog
Author(s) -
Ziegler Martin,
Soni Rohit,
Patelczyk Timo,
Ignatov Marina,
Bartsch Thorsten,
Meuffels Paul,
Kohlstedt Hermann
Publication year - 2012
Publication title -
advanced functional materials
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.069
H-Index - 322
eISSN - 1616-3028
pISSN - 1616-301X
DOI - 10.1002/adfm.201200244
Subject(s) - neuromorphic engineering , hebbian theory , associative learning , computer science , neuroscience , long term potentiation , artificial neural network , memristor , computer architecture , artificial intelligence , electronic engineering , biology , engineering , biochemistry , receptor
Neuromorphic plasticity is the basic platform for learning in biological systems and is considered as the unique concept in the brains of vertebrates, which outperform today's most powerful digital computers when it comes to cognitive and recognition tasks. An emerging task in the field of neuromorphic engineering is to mimic neural pathways via elegant technological approaches to close the gap between biological and digital computing. In this respect, functional, memristive devices are considered promising candidates with yet unknown benefit for neuromorphic circuits. It is demonstrated that a single Pt/Ge 0.3 Se 0.7 /SiO 2 /Cu memristive device implemented in an analogue circuitry mimics non‐associative and associative types of learning. For Pavlovian conditioning, different threshold voltages for the memristive device and the comparator are essential. Similarities to neurobiological correlates of learning are discussed in the framework of hebbian learning rule, plasticity, and long‐term potentiation.