An Electromagnetic Perspective of Artificial Intelligence Neuromorphic Chips DOI Creative Commons
Er‐Ping Li, Hanzhi Ma, Manareldeen Ahmed

et al.

Electromagnetic Science, Journal Year: 2023, Volume and Issue: 1(3), P. 1 - 18

Published: Sept. 1, 2023

The emergence of artificial intelligence has represented great potential in solving a wide range complex problems. However, traditional general-purpose chips based on von Neumann architectures face the “memory wall” problem when applied applications. Based efficiency human brain, many intelligent neuromorphic have been proposed to emulate its working mechanism and neuron-synapse structure. With spiking-based chips, computation energy such devices could be enhanced by integrating variety features inspired biological brain. Aligning with rapid development it is importance quickly initiate investigation electromagnetic interference signal integrity issues related for both CMOS-based memristor-based integrated circuits. Here, this paper provides review circuit design algorithms terms opportunities focus issues, modeling, optimization. Moreover, heterogeneous structures circuits other circuits, as memory arrays sensors using different integration technologies, are also reviewed, locations where might compromised discussed. Finally, we provide future trends outline prospects upcoming devices.

Language: Английский

An Electromagnetic Perspective of Artificial Intelligence Neuromorphic Chips DOI Creative Commons
Er‐Ping Li, Hanzhi Ma, Manareldeen Ahmed

et al.

Electromagnetic Science, Journal Year: 2023, Volume and Issue: 1(3), P. 1 - 18

Published: Sept. 1, 2023

The emergence of artificial intelligence has represented great potential in solving a wide range complex problems. However, traditional general-purpose chips based on von Neumann architectures face the “memory wall” problem when applied applications. Based efficiency human brain, many intelligent neuromorphic have been proposed to emulate its working mechanism and neuron-synapse structure. With spiking-based chips, computation energy such devices could be enhanced by integrating variety features inspired biological brain. Aligning with rapid development it is importance quickly initiate investigation electromagnetic interference signal integrity issues related for both CMOS-based memristor-based integrated circuits. Here, this paper provides review circuit design algorithms terms opportunities focus issues, modeling, optimization. Moreover, heterogeneous structures circuits other circuits, as memory arrays sensors using different integration technologies, are also reviewed, locations where might compromised discussed. Finally, we provide future trends outline prospects upcoming devices.

Language: Английский

Citations

13