LaMAGIC: Language-Model-based Topology Generation for Analog Integrated Circuits

Abstract

In the realm of electronic and electrical engineering, automation of analog circuit is increasingly vital given the complexity and customized require- ments of modern applications. However, existing methods only develop search-based algorithms that require many simulation iterations to design a custom circuit topology, which is usually a time-consuming process. To this end, we introduce LaMAGIC, a pioneering language model-based topology generation model that leverages super- vised finetuning for automated analog circuit design. LaMAGIC can efficiently generate an opti- mized circuit design from the custom specifica- tion in a single pass. Our approach involves a meticulous development and analysis of various input and output formulations for circuit. These formulations can ensure canonical representations of circuits and align with the autoregressive nature of LMs to effectively addressing the challenges of representing analog circuits as graphs. The experimental results show that LaMAGIC achieves a success rate of up to 96% under a strict tolerance of 0.01. We also examine the scalability and adaptability of LaMAGIC, specifically testing its performance on more complex circuits. Our findings reveal the enhanced effectiveness of our adjacency matrix-based circuit formulation with floating-point input, suggesting its suitability for handling intricate circuit designs. This research not only demonstrates the potential of language models in graph generation, but also builds a foundational framework for future explorations in automated analog circuit design.

 

Methodology
Image
method_LaMAGIC

We introduce LaMAGIC, a pioneering approach that adapts LMs to the domain of analog circuit design through SFT. This enables the efficient one-shot generation of custom circuit designs from user-defined specifications.

Results
Image
result_LaMAGIC

Experimental results show that our model achieves superior 0.96 success rate under a strict tolerance of 0.01. We further conduct an extensive evaluation of the model’s performance and its adaptability to more complex circuit designs. These evaluations provide critical insights into the effectiveness of different formulations, contributing significantly to future advancements in this field.

Citation

@misc{chang2024lamagiclanguagemodelbasedtopologygeneration,
     title={LaMAGIC: Language-Model-based Topology Generation for Analog Integrated Circuits}, 
     author={Chen-Chia Chang and Yikang Shen and Shaoze Fan and Jing Li and Shun Zhang and Ningyuan Cao and Yiran Chen and Xin Zhang},
     year={2024},
     eprint={2407.18269},
     archivePrefix={arXiv},
     primaryClass={cs.AR},
     url={https://arxiv.org/abs/2407.18269}, 
}