A Physics-Informed Recurrent Neural Network for RRAM Modeling
- Resource Type
- Conference
- Authors
- Sha, Yanliang; Lan, Jun; Li, Yida; Chen, Quan
- Source
- 2023 International Symposium of Electronics Design Automation (ISEDA) Electronics Design Automation (ISEDA), 2023 International Symposium of. :438-443 May, 2023
- Subject
- Components, Circuits, Devices and Systems
Computing and Processing
Engineering Profession
General Topics for Engineers
Photonics and Electrooptics
Recurrent neural networks
Systematics
Simulation
SPICE
Data models
Numerical models
Integrated circuit modeling
Resistive random-access memory (RRAM)
Recurrent Neural Network
Compact Model
Verilog-A
- Language
Extracting behavioural models of RRAM devices has been challenging due to their unique “memory” behaviours, for which well-established modeling frameworks and systematic parameter extraction processes have not been available. In this work, we propose a physics-informed recurrent neural network (PiRNN) methodology to generate behavioural models of RRAM devices from practical measurement/simulation data. The proposed framework can faithfully capture the evolution of internal state and its impacts on the output. A series of modifications informed by the RRAM device physics are proposed to enhance the modeling capabilities. Integration strategy with existing SPICE-type simulators is also developed. Numerical experiments with real RRAM devices data demonstrate the feasibility and advantages of the proposed methodology.