Antonio J. Peña (Barcelona Supercomputing Center, BSC) has obtained an ERC Consolidator Grant 2021. His project HomE (Enabling Homomorphic Encryption of Deep Neural Network Models and Datasets in Production Environments) will bring machine-learning users the possibility of leveraging safely their jobs with confidence in a cloud provider.

Antonio J. Peña is a Group Manager at the Barcelona Supercomputing Center (BSC), where he leads the Accelerators and Communications for High-Performance Computing Research Group. He has held prestigious named fellowships such as a Spanish Ramón y Cajal and a European Marie Sklodowska-Curie. He has also received several awards and is a Senior Member of the two major professional associations in his field: ACM and IEEE.

Antonio has been a Principal Investigator in several public and industrial projects, such as the European EPEEC, and is involved in the organization and steering committees of some of the major conferences in his field. His research interests in the area of high-performance computing include resource heterogeneity and communications.

Which are the main objectives of HomE?

This ERC grant will bring machine-learning users the possibility of leveraging their jobs with confidence in a cloud provider, even when data protection policies are in place (e.g., GDPR). Until now, homomorphically-encrypted techniques (those that enable computations directly on encrypted data) have posed unbearable data sizes and computation times. To address this, HomE will put together software optimization techniques and novel hardware designs, on top of recently-emerged memory technologies, to enable production-sized scenarios. I imagine a hospital being able to offload advanced cancer detection processes based on deep neural networks without concerns about the privacy of the medical records.

Why did you decide to carry out your research at the Barcelona Supercomputing Centre?

The Barcelona Supercomputing Centre brings me a unique opportunity to work in Spain in a top-notch centre, whose reputation is recognized worldwide. With tight industrial and academic institutional relationships and top researchers in the field, BSC is a great place to foster new ideas. I would like to highlight the great support provided by specific professionals devoted to support research, such as Project Management, Technology Transfer, Dissemination, and Communications, which really helps researchers like me to focus on our passion.

How do you expect this grant will affect your professional career?

An ERC grant is among the highest recognitions a researcher can be honoured with. I take it as a seal of quality for my activities. But now comes the time to honour the trust the ERC Scientific Council has put in me and my proposal, and deliver, and of course, ERC projects are by design highly ambitious and hence demanding. It is going to be a very intense 5-year period of my career ahead, and I expect to have a fun mixed with effort, plenty of obstacles and overcoming.

Are you planning to open research positions at your group? Which kind of profiles are you interested in?

Yes. I will be forming a team of 8 individuals specifically devoted to the ERC tasks. This will include a Sr. Researcher, two Postdoctoral Researchers, four PhD Students and a Research Engineer, covering experience in software optimization, computer architecture, and cryptography. The ERC will be a great incentive to attract more talent to my group.

Are you planning any collaborations?

There are many companies working in this area. I have a collaboration already in place with Intel on this topic, and I will be approaching IBM, NVIDIA, and Microsoft. There are teams at some of these companies working on this topic based in the UK. I am also interested in the experience in the computational algebra of the UK Science and Technology Facilities Council.

Where can we learn more about you and your project?

Group’s web page: https://www.bsc.es/discover-bsc/organisation/scientific-structure/accelerators-and-communications-hpc

People interested in collaborations and hiring opportunities may reach me at my email address: antonio.pena@bsc.es

Selected publications

Enabling homomorphically encrypted inference for large DNN models, G. Lloret-Talavera, M. Jordà, H. Servat, F. Boemer, C. Chauhan, S. Tomishima, N. N. Shah, and A. J. Peña, Transactions on Computers, IEEE (2021)

Automating the application data placement in hybrid memory systems, Servat, A. J. Peña, G. Llort, E. Mercadal, H. C. Hoppe, and J. Labarta, IEEE Cluster (2017)

During the following weeks, we will be presenting those researchers working at Spanish institutions that have obtained an ERC Starting or Consolidator Grant 2021 and their projects. You can check their profiles HERE.