You are here:

Serverless approach as a way to speed up the sensitivity analysis of computational models   

A research made by Sano scientists was among 58 (out of 280) papers accepted for the 23rd International Symposium on Cluster, Cloud and Internet Computing (CCGrid 2023). We are very proud of this success as CCGrid is a leading forum on the distributed systems topics. The Symposium took place in India, Bangalore, 1-4 May.  

The Sano paper selected for the CCGrid 2023 is devoted to a serverless approach to sensitivity analysis of computational models. Piotr Kica, Sano scientific programmer, one of the authors of the research, gave a presentation on this topic on the Symposium. 

Sensitivity analysis of computational models is an important stage in the process of performing verification and validation of medical models and simulations. Although modelling and simulation are becoming increasingly important in computational medicine, in order to bring these solutions to the market, a complex regulatory procedure needs to be followed through agencies such as the FDA in the US, or the EMA in Europe. One important stage in this process is to perform verification and validation of the models, along with the corresp onding sensitivity analysis and uncertainty quantification (VVUQ). From the computing perspective, VVUQ is a computationally intensive process, as it requires numerous runs of the model with variations of input parameters.  

The authors of the research proposed a hypothesis that serverless computing model can be a practical and efficient approach to selected cases of running VVUQ calculations. The experiments provided in the research proved that by using serverless services it is possible to acquire thousands of parallel sample computations and, as a result, speed up the calculation phase hundreds of times. 

The full list of the authors: Piotr Kica, Magdalena Otta, Krzysztof Czechowicz, Karol Zając, Piotr Nowakowski, Andrew Narracott, Ian Halliday, Maciej Malawski. 

The link to the paper: https://arxiv.org/abs/2304.08190 

Facebook
Twitter
LinkedIn