Nature: energy-efficient classification with dopant network

January 17th 2020

Prof. Peter Bobbert (TU/e, CCER) is co-author of the Nature paper Classification with a disordered dopant atom network in silicon of 15 january 2020. The paper shows that devices exploiting the correlated hopping of electrons on a network of boron dopants in silicon can perform complex classification tasks, like written character recognition. This can lead to a new energy-efficient platform for Artificial Intelligence.

 

 

Publication
Tao Chen, Jeroen van Gelder, Bram van de Ven, Sergey Amitonov, Bram de Wilde, Hans-Christian Ruiz Euler, Hajo Broersma, Peter Bobbert, Floris Zwanenburg and Wilfred van der Wiel, Classification with a disordered dopant atom network in silicon, Nature 577, 341–345 (2020). Link to paper.

Source
CCER