Title: Emerging Spintronic Device Concepts for Neuromorphic computing
Acronym of the project: CIDEXG/2022/26 VÍCTOR LÓPEZ DOMÍNGUEZ
Funding entity: Generalitat Valenciana
Modality: CIDEGENT
Entity code: CIDEXG/2022/26 VÍCTOR LÓPEZ DOMÍNGUEZ (CIDEGENT 2022)
UJI accounting code: 22I272
Principal researchers:
Total amount: 600.000,00€
Involved researchers:
Acronym of the project: CIDEXG/2022/26 VÍCTOR LÓPEZ DOMÍNGUEZ
Funding entity: Generalitat Valenciana
Modality: CIDEGENT
Entity code: CIDEXG/2022/26 VÍCTOR LÓPEZ DOMÍNGUEZ (CIDEGENT 2022)
UJI accounting code: 22I272
Principal researchers:
- Víctor López Domínguez
Total amount: 600.000,00€
Involved researchers:
- Víctor López Domínguez
Summary/Abstract:
We propose to experimentally study the electrical manipulation of ferromagnetic and antiferromagnetic materials to
design and implement the following device concepts for neuromorphic computing as follows: low energy consumption
memristors to store synaptic weights and ultra-fast neurons with inherent plasticity. Neuromorphic computing is a new
computational paradigm inspired in human brain-like computation where the data is processed in a network of neurons
connected by synapses. This computational paradigm offers a high error tolerance, parallelization, and low energy dissipation.
Traditional devices used in electronic systems, like CMOS transistors, have successfully been used to implement dedicated
algorithms to solve particular problems, but presenting as main drawback a high-power consumption, in the range of kilo
Watts and Mega Watts, low scalability due to the big number of devices need to implement neurons and synapses, and lack
of parallelization.