BSC and SimZero team up to advance industry 4.0 and smart cities with privacy-preserving, real-time scientific simulations

08 April 2024

This collaboration is a significant step forward in the digital transformation of factories and urban infrastructures, enabling intelligent and efficient advancements for a sustainable and resilient future

In a strategic collaboration to advance privacy-preserving, real-time scientific simulations on the cloud, the Barcelona Supercomputing Center - Centro Nacional de Supercomputación (BSC-CNS) is joining forces with SimZero, a deeptech startup that facilitates the use of artificial intelligence-driven scientific simulations in industry 4.0 and smart cities.

BSC is at the forefront of the European Research Council (ERC) project HomE, which focuses on making privacy-preserving machine learning (ML) viable in untrusted settings, such as cloud environments. It intends to do so using homomorphic encryption (HE), a cryptographic technique that allows computations to be performed on encrypted data without decrypting it first.

Traditionally, the computational burden of executing operations on encrypted data using HE techniques has been prohibitively high. The HomE project aims to overcome this challenge by integrating high-performance computing (HPC) software optimization strategies with innovative hardware designs, utilizing large memory pools. This approach seeks to facilitate the application of homomorphically encrypted deep learning in real-world, production-scale scenarios.

SimZero helps companies develop and deploy ML-based computational fluid dynamics (CFD) models, offering a platform that facilitates the use of artificial intelligence-driven scientific simulations. Users have the flexibility to use their own models or leverage SimZero’s open-source examples as a foundation. These models can then be uploaded to the SimZero API cloud service, facilitating effortless connectivity with digital twins, virtual reality (VR), Internet of Things (IoT), or any other software platforms.

By integrating BSC's AI and HE expertise with SimZero's advanced infrastructure for ML-CFD deployment, this collaboration offers real-time scientific models while ensuring confidentiality. Focused on experimenting with homomorphic encryption and ML-CFD, it addresses critical challenges in manufacturing, urban planning and operation of critical infrastructures. Homomorphic encryption guarantees data privacy, enabling industries and cities to benefit from AI without compromising sensitive information.

The collaboration between BSC and SimZero is a significant step forward in the digital transformation of factories and infrastructures, combining ML, CFD, and HE. This partnership aims to redefine industry standards by enabling privacy-secure, intelligent, and efficient operations across various sectors. By integrating these technologies, the collaboration enhances predictive analytics, decision-making, and data privacy, setting a new benchmark for how industries and cities can leverage the full potential of digital advancements for a sustainable and resilient future.

Antonio J. Peña, BSC researcher and Principal Investigator of the awarded ERC proposal HomE, said: “We are excited to see how HomE can empower industries to securely harness machine learning models on the cloud, safeguarding their proprietary models and sensitive data.”

SimZero’s CEO and co-founder Carlos Peña emphasized: “This advancement is not just about enhancing security; it’s about reshaping the way industries can safely adopt and scale scientific AI technologies.”