Active learning for efficient microfluidic design automation
Files
Published version
Date
2020-08-03
DOI
Authors
McIntyre, David
Lashkaripour, Ali
Densmore, Douglas
Version
Published version
OA Version
Citation
David McIntyre, Ali Lashkaripour, Douglas Densmore. 2020. "Active Learning for Efficient Microfluidic Design Automation." https://www.iwbdaconf.org/2020/docs/IWBDA2020Proceedings.pdf. 12th International Workshop on Bio-Design Automation (IWBDA-2020). Online, 2020-08-03 - 2020-08-05.
Abstract
Droplet microfluidics has the potential to eliminate the testing bottleneck in synthetic biology by screening biological samples encapsulated in water-in-oil emulsions at unprecedented throughput. Sophisticated screens require functional and complex devices that perform exactly as designed. Effective performance characterization and predictive design of droplet microfluidic components has been hampered due to low-throughput and expensive fabrication with standard soft lithography techniques. This has limited droplet microfluidics to proof-of-concept devices. Even when some of these barriers are removed through rapid prototyping, developing a robust dataset to effectively represent all parameters as a "lookup table" is near impossible.
Description
License
Attribution-NonCommercial 4.0 International