A timely update on memristors for use in neuromorphic computing.
Venema has written a lengthy and comprehensive article on developments in memristor technology for the purposes of neuromorphic computing. This is the branch of hardware computing that attempts to replicate the processing of a brain. Memristors constitute another kind of element in an electrical circuit. They combine memory with resistance. The memory feature means that, unlike transistors, they remember the state they are in even after their power has been switched off. So, goodbye random access memory. The resistance feature means that, unlike binary transistors, they possess more than two states. As such, they should be able to mimic the properties of neurons, and by extension synapses. They could hold the key to realising neuromorphic computing.
Link to article: https://www.newscientist.com/article/mg23531370-600-brain-box-multitasking-chips-that-can-match-the-human-mind/?cmpid=EMP%7CNSNS%7C2017-0608-GLOBAL-AugWk1_Memputing%7Cmemputing%20&cmpid
You may also like to browse other AI articles: http://www.thesentientrobot.com/category/ai/ai-articles/