STDP neural network unsupervised online learning circuit and method
By introducing photomemristors and long-term enhancement/suppression modules into the SNN circuit, the problem that transfer pulses and adjust synaptic weights cannot be performed simultaneously is solved, and the circuit design with lower complexity and power consumption is realized, and the integration is improved.
CN119692282AActive Publication Date: 2025-03-25UNIV OF ELECTRONICS SCI & TECH OF CHINA
7 Cites 0 Cited by
Patent Information
- Application Number
- CN202411828501.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-12
- Publication Date
- 2025-03-25
- Estimated Expiration
- 2044-12-12
AI Technical Summary
Technical Problem
Existing SNN circuits cannot be performed simultaneously when transmitting pulses and adjusting synaptic weights, resulting in increased circuit complexity and power consumption.
Method used
An unsupervised online learning circuit of STDP neural network including preneurons, posterior neurons, photomemristors, long-term suppression modules and long-term enhancement modules is adopted to perform pulse transmission and weight adjustment through photomemristors.
Benefits of technology
It is realized that the pulse transfer and synaptic weights are transferred without increasing the complexity and power consumption of the circuit, which improves the circuit integration and reduces the power consumption.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention relates to the technical field of artificial intelligence, in particular to an STDP neural network unsupervised online learning circuit and method. The circuit comprises a front neuron, a rear neuron, a photoelectric memristor, a long-time suppression module and a long-time enhancement module, and the front neuron and the rear neuron are both used for transmitting pulses; the long-time suppression module is used for reducing the weight of the photoelectric memristor; the long-time enhancement module is used for increasing the weight of the photoelectric memristor; the photoelectric memristor is used for performing pulse transmission and weight adjustment; the method comprises the following steps: comparing emission sequences of emission pulses of a front neuron and a rear neuron, and adjusting a synaptic weight of a photoelectric memristor; by introducing light regulation and control, the problem that pulse transmission and memristor weight adjustment of a traditional circuit cannot be carried out at the same time is solved; the complexity of a circuit structure and a control signal is reduced, the circuit integration level is improved, and power consumption is reduced.
Need to check novelty before this filing date? Find Prior Art