Impact Statement:All electronic design automation tools require fast, accurate and adaptive device models for efficient circuit simulation valid across various device technologies. Hence,...Show More
Abstract:
This paper presents an adaptive and automated device modeling framework valid across different technology nodes and device structures. A novel metalearning-based surrogat...Show MoreMetadata
Impact Statement:
All electronic design automation tools require fast, accurate and adaptive device models for efficient circuit simulation valid across various device technologies. Hence, this paper introduces one such groundbreaking adaptive device modeling approach by using Prior Knowledge Input with Difference Artificial Neural Network (PKID ANN) combined with Transfer Learning (TL). This PKID-TL technique allows the model to learn from existing device knowledge and adapt to new device technologies with minimal effort, drastically reducing the need for manual adjustments. This framework significantly accelerates model development process without compromising accuracy. Further, this speed improvement ultimately reduces computational burden leading to reduced CPU usage time making this approach more sustainable. Therefore, this research promises to revolutionize the semiconductor devices modeling approach, enabling speedy design cycles, less reliance on human intervention, and ultimately speeding up t...
Abstract:
This paper presents an adaptive and automated device modeling framework valid across different technology nodes and device structures. A novel metalearning-based surrogate model using Prior Knowledge Input with Difference Artificial Neural Network (PKID ANN) combined with advanced transfer learning (TL) is developed. This approach is validated using various advanced FET devices. In addition to transferring weights and biases from pretrained model, a scaled low-fidelity model is developed for efficient training of different primary target models. Two TL techniques, full and partial knowledge transfer, are compared, with PKID ANN with partial transfer learning (PKID-PTL) showing significant speed-up in all phases of model development. The proposed PKID-PTL technique is a potential candidate for efficient device modeling allowing seamless model automation across technology nodes and devices with the least human intervention.
Published in: IEEE Transactions on Artificial Intelligence ( Early Access )
