ARMM: Adaptive Reliability Quantification Model of Microfluidic Designs and Its Graph-Transformer-Based Implementation
Published in Get the paper The 42nd IEEE/ACM International Conference on Computer-Aided Design (ICCAD), 2023
Recommended citation: S.Y. Liang, M. Lian, M.C. Li, T.-M. Tseng, U. Schlichtmann, T.-Y. Ho, "ARMM: Adaptive Reliability Quantification Model of Microfluidic Designs and Its Graph-Transformer-Based Implementation," The 42nd IEEE/ACM International Conference on Computer-Aided Design (ICCAD), 2023.
In this paper, we established the first mathematical reliability model of flow-based microfluidic biochips. For a given design, the proposed models can provide reliabilities against blockage and leakage defects. Meanwhile, the model can be adapted to different manufacturing processings to make the results more useful, a detailed guide on how to do so is also providerd in the paper. In order to provide more timely feedback for designers and enable automated reliability improvement in the future, we proposed a graph-transformer-basd implementation of the mathermatical models and achieved great performance.