Synaptic Computation Enabled by Joule Heating of Single-Layered Semiconductors for Sound Localization.
Nano Lett
; 18(5): 3229-3234, 2018 05 09.
Article
em En
| MEDLINE
| ID: mdl-29668290
Synaptic computation, which is vital for information processing and decision making in neural networks, has remained technically challenging to be demonstrated without using numerous transistors and capacitors, though significant efforts have been made to emulate the biological synaptic transmission such as short-term and long-term plasticity and memory. Here, we report synaptic computation based on Joule heating and versatile doping induced metal-insulator transition in a scalable monolayer-molybdenum disulfide (MoS2) device with a biologically comparable energy consumption (â¼10 fJ). A circuit with our tunable excitatory and inhibitory synaptic devices demonstrates a key function for realizing the most precise temporal computation in the human brain, sound localization: detecting an interaural time difference by suppressing sound intensity- or frequency-dependent synaptic connectivity. This Letter opens a way to implement synaptic computing in neuromorphic applications, overcoming the limitation of scalability and power consumption in conventional CMOS-based neuromorphic devices.
Texto completo:
1
Coleções:
01-internacional
Base de dados:
MEDLINE
Idioma:
En
Revista:
Nano Lett
Ano de publicação:
2018
Tipo de documento:
Article